copy_btp_post_1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "copy_btp_post_1.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.5.2"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/sathyasravya/number-of-peacocks-automation/blob/master/copy_btp_post_1.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "g9_aGJP3tvF4",
        "trusted": true,
        "colab": {}
      },
      "source": [
        "import glob\n",
        "import os\n",
        "import librosa\n",
        "import numpy as np\n",
        "import seaborn as sns\n",
        "import wave\n",
        "import contextlib\n",
        "import pandas as pd\n",
        "\n",
        "from sklearn.metrics import accuracy_score,classification_report,f1_score\n",
        "import matplotlib.pyplot as plt\n",
        "from matplotlib.pyplot import specgram\n",
        "%matplotlib inline\n",
        "plt.style.use('ggplot')\n",
        "\n",
        "plt.rcParams['font.family'] = 'serif'\n",
        "plt.rcParams['font.serif'] = 'Ubuntu'\n",
        "plt.rcParams['font.monospace'] = 'Ubuntu Mono'\n",
        "plt.rcParams['font.size'] = 12\n",
        "plt.rcParams['axes.labelsize'] = 11\n",
        "plt.rcParams['axes.labelweight'] = 'bold'\n",
        "plt.rcParams['axes.titlesize'] = 14\n",
        "plt.rcParams['xtick.labelsize'] = 10\n",
        "plt.rcParams['ytick.labelsize'] = 10\n",
        "plt.rcParams['legend.fontsize'] = 11\n",
        "plt.rcParams['figure.titlesize'] = 13"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l4bRI1pyaDS8",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 122
        },
        "outputId": "70a489e4-3da9-4c1a-e493-052a2522c639"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3Aietf%3Awg%3Aoauth%3A2.0%3Aoob&scope=email%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdocs.test%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive.photos.readonly%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fpeopleapi.readonly&response_type=code\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "4whjGG3mZyzq",
        "colab_type": "code",
        "outputId": "3542266f-f207-477a-facc-0107fa28f2c9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "s = \"02 Eow_peacock_2.WAV\"\n",
        "s.split(\"_\")\n",
        "#print(fn.split('_')[-1]=='2.WAV')"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['02 Eow', 'peacock', '2.WAV']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "ovHpIKfEtvGG"
      },
      "source": [
        "# Features plots - Code"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "CGCfJPgjtvGM",
        "trusted": true,
        "colab": {}
      },
      "source": [
        "def load_sound_files(file_paths):\n",
        "    raw_sounds = []\n",
        "    for fp in file_paths:\n",
        "        X,sr = librosa.load(fp)\n",
        "        raw_sounds.append(X)\n",
        "    return raw_sounds\n",
        "\n",
        "def plot_waves(sound_names,raw_sounds):\n",
        "    i = 1\n",
        "    fig = plt.figure(figsize=(25,60), dpi = 900)\n",
        "    for n,f in zip(sound_names,raw_sounds):\n",
        "        plt.subplot(10,1,i)\n",
        "        librosa.display.waveplot(np.array(f),sr=22050)\n",
        "        plt.title(n.title())\n",
        "        i += 1\n",
        "    plt.suptitle('Figure 1: Waveplot',x=0.5, y=0.915,fontsize=18)\n",
        "    plt.show()\n",
        "    \n",
        "def plot_specgram(sound_names,raw_sounds):\n",
        "    i = 1\n",
        "    fig = plt.figure(figsize=(25,60), dpi = 900)\n",
        "    for n,f in zip(sound_names,raw_sounds):\n",
        "        plt.subplot(10,1,i)\n",
        "        specgram(np.array(f), Fs=22050)\n",
        "        plt.title(n.title())\n",
        "        i += 1\n",
        "    plt.suptitle('Figure 2: Spectrogram',x=0.5, y=0.915,fontsize=18)\n",
        "    plt.show()\n",
        "\n",
        "def plot_log_power_specgram(sound_names,raw_sounds):\n",
        "    i = 1\n",
        "    fig = plt.figure(figsize=(25,60), dpi = 900)\n",
        "    for n,f in zip(sound_names,raw_sounds):\n",
        "        plt.subplot(10,1,i)\n",
        "        D = librosa.logamplitude(np.abs(librosa.stft(f))**2, ref_power=np.max)\n",
        "        librosa.display.specshow(D,x_axis='time' ,y_axis='log')\n",
        "        plt.title(n.title())\n",
        "        i += 1\n",
        "    plt.suptitle('Figure 3: Log power spectrogram',x=0.5, y=0.915,fontsize=18)\n",
        "    plt.show()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "ndXOjzE0tvGT"
      },
      "source": [
        "# Extracting features"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "CRakr1AetvGW",
        "trusted": true,
        "colab": {}
      },
      "source": [
        "def extract_feature(file_name):\n",
        "    X, sample_rate = librosa.load(file_name)\n",
        "    stft = np.abs(librosa.stft(X))\n",
        "    mfccs = np.mean(librosa.feature.mfcc(y=X, sr=sample_rate, n_mfcc=40).T,axis=0)\n",
        "    chroma = np.mean(librosa.feature.chroma_stft(S=stft, sr=sample_rate).T,axis=0)\n",
        "    mel = np.mean(librosa.feature.melspectrogram(X, sr=sample_rate).T,axis=0)\n",
        "    contrast = np.mean(librosa.feature.spectral_contrast(S=stft, sr=sample_rate).T,axis=0)\n",
        "    tonnetz = np.mean(librosa.feature.tonnetz(y=librosa.effects.harmonic(X), sr=sample_rate).T,axis=0)\n",
        "    #print(chroma.shape)\n",
        "    #print(\"chroma\")\n",
        "    #print(mfccs.shape)\n",
        "    #print(\"mfccs\")\n",
        "    #print(mel.shape)\n",
        "    #print(\"mels\")\n",
        "    #print(contrast.shape)\n",
        "    #print(\"contrast\")\n",
        "    #print(tonnetz.shape)\n",
        "    #print(\"tonnetz\")    \n",
        "    return mfccs,chroma,mel,contrast,tonnetz\n",
        "\n",
        "def parse_audio_files(parent_dir,sub_dirs,file_ext='*.wav'):\n",
        "    d=0\n",
        "    duration1 = []\n",
        "    duration2 = []\n",
        "    for label, sub_dir in enumerate(sub_dirs):\n",
        "        for fn in glob.glob(os.path.join(parent_dir,sub_dir,file_ext)):\n",
        "            with contextlib.closing(wave.open(fn,'r')) as f:\n",
        "                frames = f.getnframes()\n",
        "                rate = f.getframerate()\n",
        "                duration = frames / float(rate)\n",
        "                #print(pd.Series(fn.split('_')).unique())\n",
        "                #print(d,duration)\n",
        "                d=d+1\n",
        "                #print(\" \")\n",
        "                #dn=0\n",
        "                if(sub_dir!='noise_set'):\n",
        "                    #if(duration<0):\n",
        "                    #    dn =dn+1\n",
        "                    duration1.append(duration)  \n",
        "                else:\n",
        "                    duration2.append(duration)\n",
        "            \n",
        "    return duration1,duration2"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "Qtu4lVc0Zyz4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def parse_audio_files2(parent_dir,sub_dirs):\n",
        "    features = np.empty((0,193))\n",
        "    finlabel = np.empty((0,3))\n",
        "    labells=[]\n",
        "    labells1=np.empty(0) \n",
        "    labells2=np.empty(0) \n",
        "    d=0\n",
        "    fns = []\n",
        "    labels=[]\n",
        "    label1s=[]\n",
        "    label2s=[]\n",
        "    for label, sub_dir in enumerate(sub_dirs):\n",
        "        print(sub_dir)\n",
        "        if(sub_dir =='iid' or sub_dir=='Kaan' or sub_dir =='extra_mp3'):\n",
        "          file_ext='*.WAV'\n",
        "        else:\n",
        "          file_ext='*.wav'        \n",
        "       # print(glob.glob(os.path.join(parent_dir,sub_dir,file_ext)))\n",
        "        for fn in glob.glob(os.path.join(parent_dir,sub_dir,file_ext)):\n",
        "            #print(\"heyyy0\")\n",
        "            mfccs, chroma, mel, contrast,tonnetz = extract_feature(fn)\n",
        "            ext_features = np.hstack([mfccs,chroma,mel,contrast,tonnetz])\n",
        "          #  print(\"ext_features\")\n",
        "          #  print(ext_features.shape)\n",
        "            features = np.vstack([features,ext_features])\n",
        "          #  print(\"features\")\n",
        "          #  print(features.shape)\n",
        "           # if(sub_dir != 'extra_mp3'):\n",
        "            if(sub_dir!='noise_set'):\n",
        "              #print(\"heyyy2\")\n",
        "              label =\"1\"\n",
        "              d = d+1\n",
        "              if (len(fn.split('_'))>=6):\n",
        "                label1 = \"1\"\n",
        "              else:\n",
        "                label1=\"0\"\n",
        "                label=\"1\"\n",
        "             #   print(\"heyyy3\")\n",
        "              if(sub_dir=='TRIM_old' or sub_dir=='extra_mp3'):\n",
        "                #print(\"heyyy1\")\n",
        "                if(fn.split('_')[-1]=='2.wav'or fn.split('_')[-1]=='2.WAV'):\n",
        "                  label2=\"2\"\n",
        "                if(fn.split('_')[-1]=='3.wav'or fn.split('_')[-1]=='3.WAV'):\n",
        "                  label2=\"3\"   \n",
        "                if(fn.split('_')[-1]=='1.wav'or fn.split('_')[-1]=='1.WAV'):\n",
        "                  label2=\"1\"\n",
        "                if(fn.split('_')[-1]=='MI.wav'or fn.split('_')[-1]=='MI.WAV'):\n",
        "                  label2=\"4\"\n",
        "                if(fn.split('_')[3]=='Multiplepeacock'):\n",
        "                  label2 = \"4\"\n",
        "                else:\n",
        "                  label2 =\"1\"\n",
        "              else:\n",
        "                if(fn.split('_')[-1]=='1.wav' or fn.split('_')[-1]=='1.WAV'):\n",
        "                  label2=\"1\"\n",
        "                elif(fn.split('_')[-1]=='2.wav' or fn.split('_')[-1]=='2.WAV'):\n",
        "                  label2 =\"2\" \n",
        "                elif(fn.split('_')[-1]=='3.wav' or fn.split('_')[-1]=='3.WAV'):\n",
        "                  label2 =\"3\"\n",
        "                elif(fn.split('_')[-1]=='MI.wav' or fn.split('_')[-1]=='3or4.wav' or fn.split('_')[-1]=='3or4.WAV' or fn.split('_')[-1]=='MI.WAV'):\n",
        "                  label2 =\"4\"                        \n",
        "                else:\n",
        "                  label2 =\"0\"\n",
        "                            \n",
        "            else:\n",
        "              label =\"0\"\n",
        "              label1=\"1\"\n",
        "              label2 =\"0\"\n",
        "           # print(\"heyyy\")\n",
        "            #print(label,label1,label2)\n",
        "            fns.append(fn.split('_')[1:])\n",
        "            use = np.hstack([label,label1,label2])\n",
        "            finlabel = np.vstack([finlabel,use])\n",
        "            labels.append(label)\n",
        "            label1s.append(label1)\n",
        "            label2s.append(label2)\n",
        "    return(fns, np.array(features), np.array(finlabel),np.array(labels), np.array(label1s), np.array(label2s))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3ewBs9DWZyz7",
        "colab_type": "text"
      },
      "source": [
        "> #  Feed Train & Test Data "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "yi23reLvtvGh",
        "trusted": true,
        "_kg_hide-output": true,
        "colab": {}
      },
      "source": [
        "parent_dir = '/content/drive/My Drive/Colab Notebooks/data' #/content/drive/My Drive/Colab Notebooks/data\n",
        "sub_dirs = ['iid','Kaan','C_new2','TRIM_old','A_new1','trimmed_new3','noise_set','Chincholi'] #train dirs\n",
        "sub_dirs2 = ['extra_mp3']\n",
        "duration1,duration2= parse_audio_files(parent_dir,sub_dirs)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D86MaGIrpZeQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#duration1"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "3lfrksU5Zy0A",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "dafr = pd.DataFrame()\n",
        "dafr[\"peacock_audio_durations\"] = pd.Series(duration1)\n",
        "dafr[\"noise_durations\"] = pd.Series(duration2)\n",
        "dafr.to_pickle('/content/drive/My Drive/Colab Notebooks/filedurations.pkl')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uy4n_DIzZy0D",
        "colab_type": "text"
      },
      "source": [
        "# Data Distribution"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "u-oT5HlxZy0E",
        "colab_type": "code",
        "outputId": "1d64f959-e54b-4ccb-9d54-d38c295352d9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 275
        }
      },
      "source": [
        "sns.set(style=\"whitegrid\")\n",
        "ax = sns.violinplot(data=dafr)"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAECCAYAAAAVYxsVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCCHlUHEghBBSDhUHQggh5fx/nK70eVwf9sgAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "ax6qc3R5Zy0I",
        "colab_type": "code",
        "outputId": "07585757-dd7f-4a30-f5f2-be831dd6d69b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 309
        }
      },
      "source": [
        "fn,features,final_labels,label,label1,label2 = parse_audio_files2(parent_dir,sub_dirs)"
      ],
      "execution_count": 76,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "iid\n",
            "Kaan\n",
            "C_new2\n",
            "TRIM_old\n",
            "A_new1\n",
            "trimmed_new3\n",
            "noise_set\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/librosa/core/pitch.py:145: UserWarning: Trying to estimate tuning from empty frequency set.\n",
            "  warnings.warn('Trying to estimate tuning from empty frequency set.')\n",
            "/usr/local/lib/python3.6/dist-packages/librosa/core/pitch.py:145: UserWarning: Trying to estimate tuning from empty frequency set.\n",
            "  warnings.warn('Trying to estimate tuning from empty frequency set.')\n",
            "/usr/local/lib/python3.6/dist-packages/librosa/core/pitch.py:145: UserWarning: Trying to estimate tuning from empty frequency set.\n",
            "  warnings.warn('Trying to estimate tuning from empty frequency set.')\n",
            "/usr/local/lib/python3.6/dist-packages/librosa/core/pitch.py:145: UserWarning: Trying to estimate tuning from empty frequency set.\n",
            "  warnings.warn('Trying to estimate tuning from empty frequency set.')\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Chincholi\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oP0R7I0whr7O",
        "colab_type": "text"
      },
      "source": [
        "Save and Load- Pickle files "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "INT3qzhU5b0N",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import pickle\n",
        "def feed_data_to_pkl(fn,db):\n",
        "  filename = '/content/drive/My Drive/Colab Notebooks/'+fn+'.pkl'\n",
        "  outfile = open(filename,'wb')\n",
        "  pickle.dump(db,outfile)\n",
        "  outfile.close()\n",
        "  return\n",
        "def load_data_frm_pkl(fn):\n",
        "  filename = '/content/drive/My Drive/Colab Notebooks/'+fn+'.pkl'\n",
        "  infile = open(filename,'rb')\n",
        "  new_dict = pickle.load(infile)\n",
        "  infile.close()\n",
        "  return new_dict"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bczQIcUzldJY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "db1 = {}\n",
        "db2 ={}\n",
        "db1 = {'filename': fn, 'features': features, 'fin_labels':final_labels, 'label':label,\n",
        "       'label1':label1,'label2':label2 } \n",
        "filename = 'good_data'\n",
        "feed_data_to_pkl(filename,db1)\n",
        "#outfile = open(filename,'wb')\n",
        "#pickle.dump(db1,outfile)\n",
        "#outfile.close()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b0q1i1Rzqj8p",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "new_dict = load_data_frm_pkl('good_data')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ffSq0xoXskKK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "fn = new_dict['filename']\n",
        "features = new_dict['features']\n",
        "final_labels =new_dict['fin_labels'] \n",
        "label = new_dict['label'] \n",
        "label1 = new_dict['label1']\n",
        "label2 = new_dict['label2']"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hE8SZsI4u18g",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#pd.Series(features).shape"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ti3ohgNpsNuG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def to_df(fn,label,label1,label2):\n",
        "  d = pd.DataFrame()\n",
        "#pd.DataFrame(list(new_dict.items()),columns=['filename', 'features','fin_labels','label','label1','label2'])\n",
        "  d['filename'] = pd.Series(fn)\n",
        "  #d['features'] =pd.Series(features)\n",
        "  #d['fin_labels'] = final_labels\n",
        "  d['label'] = label\n",
        "  d['label1'] = label1\n",
        "  d['label2'] = label2\n",
        "  return d\n",
        "d = to_df(fn,label,label1,label2)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ENGMWKiNvL4Z",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "d.to_csv('/content/drive/My Drive/Colab Notebooks/data1.csv')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e6dPP-_3q3cX",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "514ebd8c-cbe4-4209-e898-05d0c45d53dd"
      },
      "source": [
        "label2.shape"
      ],
      "execution_count": 84,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1272,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 84
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fm5kN3E-zgHI",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "c65c274a-d912-4a03-a7da-fe0606940313"
      },
      "source": [
        "fn_new,features_new,fl_new,label_new,label1_new,label2_new = parse_audio_files2(parent_dir,sub_dirs2)"
      ],
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "extra_mp3\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "oQuhLt3lZy0L",
        "colab_type": "code",
        "outputId": "b67b4a2b-05db-4ee3-aa22-5e5cc5b344f4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "print(fl_new.shape)\n",
        "final_labels.shape"
      ],
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(30, 3)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1272, 3)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3ImH8dKNrgdE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "##filename = 'train_feat_labels'\n",
        "db2={}\n",
        "db2 = {'filename': fn_new, 'features': features_new, 'fin_labels':fl_new, 'label':label_new,\n",
        "      'label1':label1_new,'label2':label2_new }\n",
        "file2 = 'good_data2'\n",
        "feed_data_to_pkl(file2,db2)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "shIwlifrw5lQ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "6ca3151a-fac1-4345-93b3-c7f794446e26"
      },
      "source": [
        "pd.Series(db2['filename']).shape"
      ],
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(30,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RXzxenesvfLC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "d2 = to_df(fn_new,label_new,label1_new,label2_new)\n",
        "d2.to_csv('/content/drive/My Drive/Colab Notebooks/data2.csv')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g4O4wo1_wBAE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "new_dict2 = load_data_frm_pkl('good_data2')\n",
        "fn_new = new_dict2['filename']\n",
        "features_new = new_dict2['features']\n",
        "fl_new =new_dict2['fin_labels'] \n",
        "label_new = new_dict2['label'] \n",
        "label1_new = new_dict2['label1']\n",
        "label2_new = new_dict2['label2']"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ijvhrheewXoi",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "889d6c0f-e610-409f-e31f-769fc3464064"
      },
      "source": [
        "fl_new.shape"
      ],
      "execution_count": 85,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(30, 3)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 85
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1mRAA6iNr6VB",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#filename = '/content/drive/My Drive/Colab Notebooks/tr_fea_labels.pkl'\n",
        "#infile = open(filename,'rb')\n",
        "#new_dict = pickle.load(infile)\n",
        "#infile.close()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fGL27HFbt-LT",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "1daa1d41-e75b-42f6-e8fe-ada07c68b26f"
      },
      "source": [
        "features.shape"
      ],
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1272, 193)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 86
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DwAEsHLWlgxV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MUhyEoQXk51p",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#outfile = open(file2,'wb')\n",
        "#pickle.dump(db2,outfile)\n",
        "#outfile.close()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "5Mt6VbZPZy0O",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.preprocessing import MultiLabelBinarizer,LabelBinarizer\n",
        "from numpy import argmax\n",
        "from keras.utils import to_categorical\n",
        "from keras import initializers\n",
        "\n",
        "one_hot1 = LabelBinarizer()\n",
        "# One-hot encode data\n",
        "label_ = one_hot1.fit_transform(label)\n",
        "label_1 = one_hot1.fit_transform(label1)\n",
        "#print(one_hot1.classes_)print(label_)print(label_1)\n",
        "one_hot2 = MultiLabelBinarizer()\n",
        "# One-hot encode data\n",
        "#label_2 = one_hot2.fit_transform(label2)\n",
        "label__2 = to_categorical(label2)\n",
        "#print(label__2)\n",
        "#one_hot2.classes_"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "etZQcctOZy0S",
        "colab_type": "code",
        "outputId": "cf7a3641-8f6e-41d4-9b4b-a2b8791dc450",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "print(label__2.shape)\n",
        "print(label_1.shape)\n",
        "print(label_.shape)\n",
        "#print(label_2.shape)"
      ],
      "execution_count": 88,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(1272, 5)\n",
            "(1272, 1)\n",
            "(1272, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "5XNccvKWZy0a",
        "colab_type": "code",
        "outputId": "11c5ab35-7c49-4786-e172-3fdc94a39644",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 779
        }
      },
      "source": [
        "def show_bar_dist_data(d1):\n",
        "  d1['label'].value_counts().plot(kind='bar')\n",
        "            #,'disturbance','single_peacock']].plot(kind='bar')\n",
        "#plt.xticks(np.arange(0, 51, 5)) \n",
        "  plt.yticks(np.arange(0, 1200, 50)) \n",
        "  plt.show()\n",
        "  d1['label1'].value_counts().plot(kind='bar')\n",
        "  plt.yticks(np.arange(0, 1200, 50)) \n",
        "  plt.show()\n",
        "  d1['label2'].value_counts().plot(kind='bar')\n",
        "  plt.yticks(np.arange(0, 1200, 25)) \n",
        "  plt.show()\n",
        "show_bar_dist_data(d)"
      ],
      "execution_count": 89,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAD+CAYAAAAgT5JOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAD+CAYAAAAgT5JOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBAAgJgYEACCm/wfkI/ncQpMLRgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAD+CAYAAAAgT5JOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nAh2pLCmyu5N",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 779
        },
        "outputId": "3ced634b-3c0a-4c00-8862-28043b11c0a2"
      },
      "source": [
        "show_bar_dist_data(d2)"
      ],
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAD+CAYAAAAgT5JOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n7qyfIx0AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAD+CAYAAAAgT5JOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nigAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAD+CAYAAAAgT5JOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIYQQ4qakQAghhLip/wcsVQgzu+YDIAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "_WU4ITASZy0d",
        "colab_type": "code",
        "outputId": "32ae8813-404c-41fb-c337-4993d3f03b62",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "pd.Series(pd.Series(label).value_counts())"
      ],
      "execution_count": 138,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0    1143\n",
              "1     129\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 138
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "riVZ7-VCZy0g",
        "colab_type": "code",
        "outputId": "a9393150-73f0-426e-fba7-c43251f6efda",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "pd.Series(pd.Series(label1).value_counts())"
      ],
      "execution_count": 139,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1    1174\n",
              "0      98\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 139
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6ryOH1IHSPP6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def return_median(gg):\n",
        "  g=(np.array(gg))\n",
        "  g.sort()\n",
        "  n = len(g)\n",
        "  print(n)  \n",
        "  if(n%2==0):\n",
        "    x= (g[int(n/2)]+g[int((n/2) -1)])/2\n",
        "  else:\n",
        "    x= g[int((n-1)/2)] \n",
        "  return int(x)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T4AcJ_yXSjXQ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "20da9080-c090-4eb7-d7d8-bd1a2ec02b67"
      },
      "source": [
        "print(return_median([1,2,3,4]))"
      ],
      "execution_count": 174,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "4\n",
            "2\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "l6qMNHkEZy0k",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "gg=(pd.Series(pd.Series(label2).value_counts()))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "Aj-3YAWttvGr"
      },
      "source": [
        "# MLP keras "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Kgb0pURqtvGt",
        "trusted": true,
        "colab": {}
      },
      "source": [
        "import keras \n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Activation,Flatten, Dropout\n",
        "from keras.layers.normalization import BatchNormalization\n",
        "\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "from sklearn.model_selection import train_test_split\n",
        "from keras.layers import Conv2D, MaxPooling2D\n",
        "#from keras import regularizers, optimizer\n",
        "import pandas as pd\n",
        "import random\n",
        "import tensorflow as tf\n",
        "from sklearn.metrics import confusion_matrix,precision_score,recall_score,f1_score,multilabel_confusion_matrix\n",
        "from keras.wrappers.scikit_learn import KerasClassifier\n",
        "from keras.utils import np_utils\n",
        "from keras.layers import Input\n",
        "from keras.models import Model\n",
        "from keras.optimizers import Adam\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import KFold\n",
        "from keras import backend as K\n",
        "random.seed(42)\n",
        "np.random.seed(42)\n",
        "tf.set_random_seed(42)\n",
        "\n",
        "#my_init = initializers.glorot_uniform(seed=42)\n",
        "#session_conf = tf.ConfigProto(intra_op_parallelism_threads=1,\n",
        "#                              inter_op_parallelism_threads=1)\n",
        "#sess = tf.Session(graph=tf.get_default_graph(), config=session_conf)\n",
        "#K.set_session(sess)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BJj2uvm3PxYt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "KUWHsjfmZy1A",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def f1(y_true, y_pred):\n",
        "    y_pred = K.round(y_pred)\n",
        "    tp = K.sum(K.cast(y_true*y_pred, 'float'), axis=0)\n",
        "    tn = K.sum(K.cast((1-y_true)*(1-y_pred), 'float'), axis=0)\n",
        "    fp = K.sum(K.cast((1-y_true)*y_pred, 'float'), axis=0)\n",
        "    fn = K.sum(K.cast(y_true*(1-y_pred), 'float'), axis=0)\n",
        "\n",
        "    p = tp / (tp + fp + K.epsilon())\n",
        "    r = tp / (tp + fn + K.epsilon())\n",
        "\n",
        "    f1 = 2*p*r / (p+r+K.epsilon())\n",
        "    f1 = tf.where(tf.is_nan(f1), tf.zeros_like(f1), f1)\n",
        "    return K.mean(f1)\n",
        "\n",
        "def f1_loss(y_true, y_pred):\n",
        "    \n",
        "    tp = K.sum(K.cast(y_true*y_pred, 'float'), axis=0)\n",
        "    tn = K.sum(K.cast((1-y_true)*(1-y_pred), 'float'), axis=0)\n",
        "    fp = K.sum(K.cast((1-y_true)*y_pred, 'float'), axis=0)\n",
        "    fn = K.sum(K.cast(y_true*(1-y_pred), 'float'), axis=0)\n",
        "\n",
        "    p = tp / (tp + fp + K.epsilon())\n",
        "    r = tp / (tp + fn + K.epsilon())\n",
        "\n",
        "    f1 = 2*p*r / (p+r+K.epsilon())\n",
        "    f1 = tf.where(tf.is_nan(f1), tf.zeros_like(f1), f1)\n",
        "    return 1 - K.mean(f1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "Uv2grtcVZy0-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def recall_m(y_true, y_pred):\n",
        "        true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))\n",
        "        possible_positives = K.sum(K.round(K.clip(y_true, 0, 1)))\n",
        "        recall = true_positives / (possible_positives + K.epsilon())\n",
        "        return recall\n",
        "\n",
        "def precision_m(y_true, y_pred):\n",
        "        true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))\n",
        "        predicted_positives = K.sum(K.round(K.clip(y_pred, 0, 1)))\n",
        "        precision = true_positives / (predicted_positives + K.epsilon())\n",
        "        return precision\n",
        "\n",
        "def f1_m(y_true, y_pred):\n",
        "    precision = precision_m(y_true, y_pred)\n",
        "    recall = recall_m(y_true, y_pred)\n",
        "    return 2*((precision*recall)/(precision+recall+K.epsilon()))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "adfAzJz-uSwA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import time\n",
        "from functools import wraps\n",
        "def timeit(f):\n",
        "    @wraps(f)\n",
        "    def wrapper(*args, **kwds):\n",
        "        start_time = time.time()\n",
        "        result = f(*args, **kwds)\n",
        "        elapsed_time = time.time() - start_time\n",
        "        print('Elapsed computation time: {:.3f} secs'\n",
        "              .format(elapsed_time))\n",
        "        return (elapsed_time, result)\n",
        "    return wrapper"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "OYlHjiYWtvG1",
        "trusted": true,
        "colab": {}
      },
      "source": [
        "X_train, X_test, y1_train, y1_test,y2_train,y2_test,y3_train,y3_test = train_test_split(features,\n",
        "                                            label_,label_1,label__2, test_size=0.33, random_state=42)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mYDIKWa_Ew8x",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "db_train_test ={}\n",
        "db_train_test = { 'X_train':X_train,'y1_train':y1_train,'y2_train':y2_train,'y3_train':y3_train,\n",
        "                 'X_test':X_test,'y1_test':y1_test,'y2_test':y2_test,'y3_test':y3_test}\n",
        "file_tr_te = 'data_train_test'\n",
        "feed_data_to_pkl(file_tr_te,db_train_test)\n",
        "#db_train_test['X_train'] = X_train\n",
        "#db_train_test['y1_train']=y1_train\n",
        "#db_train_test['y1_test'] = y1_test\n",
        "#db_train_test['y2_train'] = y2_train\n",
        "#db_train_test['y2_test'] =y2_test\n",
        "#db_train_test['y3_test'] = y3_test\n",
        "#db_train_test['y3_train'] = y3_train\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jEUt-tMoOyJi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "0lzGtpAMtvHE"
      },
      "source": [
        "## Multi output neural network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "uC4BdgmaZy1D",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def make_model(n_dim):\n",
        "  #create the model\n",
        "  inn=(Input(shape = (n_dim,)))\n",
        "  x= (Sequential())(inn)\n",
        "  # Adding the input layer and the first hidden layer\n",
        "  x=(Dense(output_dim = 40, init = 'uniform', activation = 'relu', input_dim = 193))(x)\n",
        "  x=(BatchNormalization())(x)\n",
        "  x=(Dropout(0.2))(x)\n",
        "  # Adding the second hidden layer\n",
        "  x=(Dense(output_dim = 20, init = 'uniform', activation = 'relu'))(x)\n",
        "  x=(BatchNormalization())(x)\n",
        "  x=(Dropout(0.2))(x)\n",
        "\n",
        "  x=(Dense(output_dim = 8, init = 'uniform', activation = 'relu'))(x)\n",
        "  x=(BatchNormalization())(x)\n",
        "  x=(Dropout(0.2))(x)\n",
        "  # Adding the output layer\n",
        "  out1 =(Dense(output_dim = 1, init = 'uniform', activation = 'sigmoid'))(x)\n",
        "  out2 =(Dense(output_dim = 1, init = 'uniform', activation = 'sigmoid'))(x)\n",
        "  out3 =(Dense(output_dim = 5, init = 'uniform', activation = 'softmax'))(x)\n",
        "  # Compiling Neural Network\n",
        "  model = Model(inputs=[inn], outputs=[out1, out2,out3])\n",
        "  #classifier.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])\n",
        "  # compile the model\n",
        "  model.compile(optimizer='adam', loss=['binary_crossentropy','binary_crossentropy','categorical_crossentropy'],\n",
        "              metrics=[f1_m,f1_m,f1_m])\n",
        "  return model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HibhP_-Iirls",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def make_model_(n_dim):\n",
        "  #create the model\n",
        "  inn=(Input(shape = (n_dim,)))\n",
        "  x= (Sequential())(inn)\n",
        "  # Adding the input layer and the first hidden layer\n",
        "  x=(Dense(output_dim = 40, init = 'uniform', activation = 'relu', input_dim = 193))(x)\n",
        "  x=(BatchNormalization())(x)\n",
        "  x=(Dropout(0.2))(x)\n",
        "  # Adding the second hidden layer\n",
        "  x=(Dense(output_dim = 20, init = 'uniform', activation = 'relu'))(x)\n",
        "  x=(BatchNormalization())(x)\n",
        "  x=(Dropout(0.2))(x)\n",
        "\n",
        "  x=(Dense(output_dim = 8, init = 'uniform', activation = 'relu'))(x)\n",
        "  x=(BatchNormalization())(x)\n",
        "  x=(Dropout(0.2))(x)\n",
        "  # Adding the output layer\n",
        "  out1 =(Dense(output_dim = 1, init = 'uniform', activation = 'sigmoid'))(x)\n",
        "  #out2 =(Dense(output_dim = 1, init = 'uniform', activation = 'sigmoid'))(x)\n",
        "  #out3 =(Dense(output_dim = 5, init = 'uniform', activation = 'softmax'))(x)\n",
        "  # Compiling Neural Network\n",
        "  model = Model(inputs=[inn], outputs=[out1])\n",
        "  #classifier.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])\n",
        "  # compile the model\n",
        "  #,'binary_crossentropy','categorical_crossentropy'],\n",
        "  model.compile(optimizer='adam', loss=['binary_crossentropy'],\n",
        "              metrics=[f1_m])\n",
        "  return model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fE669_LLi1XW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def make_model_2(n_dim):\n",
        "  #create the model\n",
        "  inn=(Input(shape = (n_dim,)))\n",
        "  x= (Sequential())(inn)\n",
        "  # Adding the input layer and the first hidden layer\n",
        "  x=(Dense(output_dim = 40, init = 'uniform', activation = 'relu', input_dim = 193))(x)\n",
        "  x=(BatchNormalization())(x)\n",
        "  x=(Dropout(0.2))(x)\n",
        "  # Adding the second hidden layer\n",
        "  x=(Dense(output_dim = 20, init = 'uniform', activation = 'relu'))(x)\n",
        "  x=(BatchNormalization())(x)\n",
        "  x=(Dropout(0.2))(x)\n",
        "\n",
        "  x=(Dense(output_dim = 8, init = 'uniform', activation = 'relu'))(x)\n",
        "  x=(BatchNormalization())(x)\n",
        "  x=(Dropout(0.2))(x)\n",
        "  # Adding the output layer\n",
        "  #out1 =(Dense(output_dim = 1, init = 'uniform', activation = 'sigmoid'))(x)\n",
        "  #out2 =(Dense(output_dim = 1, init = 'uniform', activation = 'sigmoid'))(x)\n",
        "  out3 =(Dense(output_dim = 5, init = 'uniform', activation = 'softmax'))(x)\n",
        "  # Compiling Neural Network\n",
        "  model = Model(inputs=[inn], outputs=[out3])\n",
        "  #classifier.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])\n",
        "  # compile the model\n",
        "  model.compile(optimizer='adam', loss=['categorical_crossentropy'],\n",
        "              metrics=[f1_m])\n",
        "  return model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yWnYTDfl9L8R",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 88
        },
        "outputId": "f6c6e327-7106-469e-ae6e-cecd7e90f8a4"
      },
      "source": [
        "from imblearn.keras import BalancedBatchGenerator\n",
        "def fit_predict_balanced_model(X_train, y1_train,y2_train,y3_train, X_test, y1_test,y2_test,y3_test):\n",
        "    model1 = make_model_(X_train.shape[1])\n",
        "    #print(model.summary())\n",
        "    training_generator = BalancedBatchGenerator(X_train,y1_train,\n",
        "                                                batch_size=30,\n",
        "                                                random_state=42)\n",
        "    history1 = model1.fit_generator(generator=training_generator,epochs=50, verbose=1)\n",
        "    yy1= model1.predict(X_test, batch_size=30)\n",
        "    model2 = make_model_(X_train.shape[1])\n",
        "    training_generator = BalancedBatchGenerator(X_train,y2_train,\n",
        "                                                batch_size=30,\n",
        "                                                random_state=42)\n",
        "    history2 = model2.fit_generator(generator=training_generator, epochs=50, verbose=1)\n",
        "    yy2= model2.predict(X_test, batch_size=30)\n",
        "    model3 = make_model_2(X_train.shape[1])\n",
        "    training_generator = BalancedBatchGenerator(X_train,y3_train,\n",
        "                                                            random_state=42)\n",
        "    history3 = model2.fit_generator(generator=training_generator, epochs=10, verbose=1)\n",
        "    yy3= model3.predict(X_test)\n",
        "    return yy1,yy2,yy3,history1,history2,history3,model1,model2,model3"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/externals/six.py:31: DeprecationWarning: The module is deprecated in version 0.21 and will be removed in version 0.23 since we've dropped support for Python 2.7. Please rely on the official version of six (https://pypi.org/project/six/).\n",
            "  \"(https://pypi.org/project/six/).\", DeprecationWarning)\n",
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HYW8HyafnG3y",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "06ad3f56-fe12-4148-ca5d-cc356621d974"
      },
      "source": [
        "y3_train.shape"
      ],
      "execution_count": 194,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(852, 5)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 194
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "r0wY3LqKkbkM",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 200
        },
        "outputId": "1b35ab7f-b315-4afa-ed70-1c9cf8e6a1cc"
      },
      "source": [
        " yy1_b,yy2_b,yy3_b,history1_b,history2_b,history3_b,model1_b,model2_b,model3_b= fit_predict_balanced_model(X_train,\n",
        "                              y1_train,y2_train,y3_train, X_test, y1_test,y2_test,y3_test)"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-2-f78821d2275c>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m yy1_b,yy2_b,yy3_b,history1_b,history2_b,history3_b,model1_b,model2_b,model3_b= fit_predict_balanced_model(X_train,\n\u001b[0m\u001b[1;32m      2\u001b[0m                              y1_train,y2_train,y3_train, X_test, y1_test,y2_test,y3_test)\n",
            "\u001b[0;31mNameError\u001b[0m: name 'X_train' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "If1kUtvP8zM4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import roc_auc_score\n",
        "def fit_predict_imbalanced_model(X_train, y1_train,y2_train,y3_train, X_test, y1_test,y2_test,y3_test):\n",
        "    model = make_model(X_train.shape[1])\n",
        "    print(model.summary())\n",
        "    history = model.fit(X_train, [y1_train,y2_train,y3_train],validation_split=0.3,epochs=50, verbose=1, batch_size=30)\n",
        "    yy1,yy2,yy3 = model.predict(X_test)\n",
        "    print(yy1.shape)\n",
        "    print(yy2.shape)\n",
        "    print(yy3.shape)\n",
        "    return yy1,yy2,yy3,history,model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DvT-8XW_Pldl",
        "colab_type": "text"
      },
      "source": [
        "Class weights"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bBeAWsBIQjs5",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "0ddac2de-34f2-4c02-a3d8-95a55049be32"
      },
      "source": [
        "x1 = pd.Series(pd.Series(label).value_counts())\n",
        "x2 = pd.Series(pd.Series(label1).value_counts())\n",
        "x3 = pd.Series(pd.Series(label2).value_counts())\n",
        "c1=return_median(x1)\n",
        "c2 = return_median(x2)\n",
        "c3 =return_median(x3)\n",
        "class_weights = {0: c3/x3[0],\n",
        "                1: c3/x3[1],\n",
        "                2: c3/x3[2],\n",
        "                3: c3/x3[3],\n",
        "                4: c3/x3[4]} # w(i) = N/n(i)\n",
        "cw1 = {0: c1/x1[0],\n",
        "       1: c1/x1[1]} \n",
        "cw2 = {0: c2/x2[0],\n",
        "       1: c2/x2[1]} \n",
        "def fit_predict_weighted_model(X_train, y1_train,y2_train,y3_train, X_test, y1_test,\n",
        "                               y2_test,y3_test,cw1,cw2,class_weights):\n",
        "    model = make_model(X_train.shape[1])\n",
        "    #print(model.summary())\n",
        "    \n",
        "    history = model.fit(X_train, [y1_train,y2_train,y3_train],validation_split=0.3,epochs=50, \n",
        "                        class_weight=[cw1,cw2,class_weights],verbose=1, batch_size=30)\n",
        "    yy1,yy2,yy3 = model.predict(X_test)\n",
        "    print(yy1.shape)\n",
        "    print(yy2.shape)\n",
        "    print(yy3.shape)\n",
        "    return yy1,yy2,yy3,history,model"
      ],
      "execution_count": 176,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2\n",
            "2\n",
            "5\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "_dkHVaj8Zy1J",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "85273370-6b82-4d6a-8dae-5de1e80d8b2c"
      },
      "source": [
        "print(cw1,cw2)\n",
        "class_weights"
      ],
      "execution_count": 179,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "{0: 0.5564304461942258, 1: 4.930232558139535} {0: 0.5417376490630323, 1: 6.489795918367347}\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{0: 0.041884816753926704, 1: 0.75, 2: 1.0, 3: 4.363636363636363, 4: 16.0}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 179
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5ZDpQjsUA5Ba",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "yy1,yy2,yy3,his_plain,model_plain = fit_predict_imbalanced_model(X_train, y1_train,y2_train,\n",
        "                                                     y3_train, X_test, y1_test,y2_test,y3_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uazFkBtIIJsX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "db_plain_res = {}\n",
        "db_plain_res = {'yy1':yy1,'yy2':yy2,'yy3':yy3,'history':his_plain}\n",
        "feed_data_to_pkl('plain_res_his',db_plain_res)\n",
        "feed_data_to_pkl('plain_model',model_plain)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "K8T3iqXoWkMm",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "1e4e6df3-357d-49f3-e659-7cd3f3152451"
      },
      "source": [
        "yy1_w,yy2_w,yy3_w,his_w,model_w = fit_predict_weighted_model(X_train, y1_train,y2_train,\n",
        "                                                     y3_train, X_test, y1_test,y2_test,y3_test,cw1,cw2,class_weights)"
      ],
      "execution_count": 181,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:6: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(activation=\"relu\", input_dim=193, units=40, kernel_initializer=\"uniform\")`\n",
            "  \n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:10: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(activation=\"relu\", units=20, kernel_initializer=\"uniform\")`\n",
            "  # Remove the CWD from sys.path while we load stuff.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:14: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(activation=\"relu\", units=8, kernel_initializer=\"uniform\")`\n",
            "  \n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:18: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(activation=\"sigmoid\", units=1, kernel_initializer=\"uniform\")`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:19: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(activation=\"sigmoid\", units=1, kernel_initializer=\"uniform\")`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:20: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(activation=\"softmax\", units=5, kernel_initializer=\"uniform\")`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Train on 596 samples, validate on 256 samples\n",
            "Epoch 1/50\n",
            "596/596 [==============================] - 5s 8ms/step - loss: 5.2168 - dense_76_loss: 0.7097 - dense_77_loss: 4.0890 - dense_78_loss: 0.4181 - dense_76_f1_m: 0.1704 - dense_76_f1_m_1: 0.1704 - dense_76_f1_m_2: 0.1704 - dense_77_f1_m: 0.5925 - dense_77_f1_m_1: 0.5925 - dense_77_f1_m_2: 0.5925 - dense_78_f1_m: 0.0000e+00 - dense_78_f1_m_1: 0.0000e+00 - dense_78_f1_m_2: 0.0000e+00 - val_loss: 5.3997 - val_dense_76_loss: 0.6256 - val_dense_77_loss: 4.2097 - val_dense_78_loss: 0.5644 - val_dense_76_f1_m: 0.1784 - val_dense_76_f1_m_1: 0.1784 - val_dense_76_f1_m_2: 0.1784 - val_dense_77_f1_m: 0.6043 - val_dense_77_f1_m_1: 0.6043 - val_dense_77_f1_m_2: 0.6043 - val_dense_78_f1_m: 0.0000e+00 - val_dense_78_f1_m_1: 0.0000e+00 - val_dense_78_f1_m_2: 0.0000e+00\n",
            "Epoch 2/50\n",
            "596/596 [==============================] - 0s 212us/step - loss: 5.1239 - dense_76_loss: 0.6921 - dense_77_loss: 4.0183 - dense_78_loss: 0.4136 - dense_76_f1_m: 0.2323 - dense_76_f1_m_1: 0.2323 - dense_76_f1_m_2: 0.2323 - dense_77_f1_m: 0.7171 - dense_77_f1_m_1: 0.7171 - dense_77_f1_m_2: 0.7171 - dense_78_f1_m: 0.0000e+00 - dense_78_f1_m_1: 0.0000e+00 - dense_78_f1_m_2: 0.0000e+00 - val_loss: 5.1887 - val_dense_76_loss: 0.6057 - val_dense_77_loss: 4.0317 - val_dense_78_loss: 0.5514 - val_dense_76_f1_m: 0.2268 - val_dense_76_f1_m_1: 0.2268 - val_dense_76_f1_m_2: 0.2268 - val_dense_77_f1_m: 0.7893 - val_dense_77_f1_m_1: 0.7893 - val_dense_77_f1_m_2: 0.7893 - val_dense_78_f1_m: 0.0000e+00 - val_dense_78_f1_m_1: 0.0000e+00 - val_dense_78_f1_m_2: 0.0000e+00\n",
            "Epoch 3/50\n",
            "596/596 [==============================] - 0s 161us/step - loss: 5.0171 - dense_76_loss: 0.6738 - dense_77_loss: 3.9403 - dense_78_loss: 0.4031 - dense_76_f1_m: 0.2693 - dense_76_f1_m_1: 0.2693 - dense_76_f1_m_2: 0.2693 - dense_77_f1_m: 0.7401 - dense_77_f1_m_1: 0.7401 - dense_77_f1_m_2: 0.7401 - dense_78_f1_m: 0.0000e+00 - dense_78_f1_m_1: 0.0000e+00 - dense_78_f1_m_2: 0.0000e+00 - val_loss: 5.0132 - val_dense_76_loss: 0.5839 - val_dense_77_loss: 3.8946 - val_dense_78_loss: 0.5347 - val_dense_76_f1_m: 0.2679 - val_dense_76_f1_m_1: 0.2679 - val_dense_76_f1_m_2: 0.2679 - val_dense_77_f1_m: 0.7840 - val_dense_77_f1_m_1: 0.7840 - val_dense_77_f1_m_2: 0.7840 - val_dense_78_f1_m: 0.0000e+00 - val_dense_78_f1_m_1: 0.0000e+00 - val_dense_78_f1_m_2: 0.0000e+00\n",
            "Epoch 4/50\n",
            "596/596 [==============================] - 0s 182us/step - loss: 4.8535 - dense_76_loss: 0.6418 - dense_77_loss: 3.8202 - dense_78_loss: 0.3914 - dense_76_f1_m: 0.3155 - dense_76_f1_m_1: 0.3155 - dense_76_f1_m_2: 0.3155 - dense_77_f1_m: 0.7940 - dense_77_f1_m_1: 0.7940 - dense_77_f1_m_2: 0.7940 - dense_78_f1_m: 0.0000e+00 - dense_78_f1_m_1: 0.0000e+00 - dense_78_f1_m_2: 0.0000e+00 - val_loss: 4.5628 - val_dense_76_loss: 0.5432 - val_dense_77_loss: 3.5005 - val_dense_78_loss: 0.5191 - val_dense_76_f1_m: 0.4949 - val_dense_76_f1_m_1: 0.4949 - val_dense_76_f1_m_2: 0.4949 - val_dense_77_f1_m: 0.9520 - val_dense_77_f1_m_1: 0.9520 - val_dense_77_f1_m_2: 0.9520 - val_dense_78_f1_m: 0.0000e+00 - val_dense_78_f1_m_1: 0.0000e+00 - val_dense_78_f1_m_2: 0.0000e+00\n",
            "Epoch 5/50\n",
            "596/596 [==============================] - 0s 164us/step - loss: 4.7065 - dense_76_loss: 0.6102 - dense_77_loss: 3.7129 - dense_78_loss: 0.3834 - dense_76_f1_m: 0.3495 - dense_76_f1_m_1: 0.3495 - dense_76_f1_m_2: 0.3495 - dense_77_f1_m: 0.8278 - dense_77_f1_m_1: 0.8278 - dense_77_f1_m_2: 0.8278 - dense_78_f1_m: 0.0000e+00 - dense_78_f1_m_1: 0.0000e+00 - dense_78_f1_m_2: 0.0000e+00 - val_loss: 4.2041 - val_dense_76_loss: 0.5080 - val_dense_77_loss: 3.1852 - val_dense_78_loss: 0.5109 - val_dense_76_f1_m: 0.8089 - val_dense_76_f1_m_1: 0.8089 - val_dense_76_f1_m_2: 0.8089 - val_dense_77_f1_m: 0.9769 - val_dense_77_f1_m_1: 0.9769 - val_dense_77_f1_m_2: 0.9769 - val_dense_78_f1_m: 0.0000e+00 - val_dense_78_f1_m_1: 0.0000e+00 - val_dense_78_f1_m_2: 0.0000e+00\n",
            "Epoch 6/50\n",
            "596/596 [==============================] - 0s 169us/step - loss: 4.5227 - dense_76_loss: 0.5637 - dense_77_loss: 3.5924 - dense_78_loss: 0.3666 - dense_76_f1_m: 0.3767 - dense_76_f1_m_1: 0.3767 - dense_76_f1_m_2: 0.3767 - dense_77_f1_m: 0.8476 - dense_77_f1_m_1: 0.8476 - dense_77_f1_m_2: 0.8476 - dense_78_f1_m: 0.0000e+00 - dense_78_f1_m_1: 0.0000e+00 - dense_78_f1_m_2: 0.0000e+00 - val_loss: 3.2343 - val_dense_76_loss: 0.5284 - val_dense_77_loss: 2.1587 - val_dense_78_loss: 0.5473 - val_dense_76_f1_m: 0.1211 - val_dense_76_f1_m_1: 0.1211 - val_dense_76_f1_m_2: 0.1211 - val_dense_77_f1_m: 0.9696 - val_dense_77_f1_m_1: 0.9696 - val_dense_77_f1_m_2: 0.9696 - val_dense_78_f1_m: 0.0000e+00 - val_dense_78_f1_m_1: 0.0000e+00 - val_dense_78_f1_m_2: 0.0000e+00\n",
            "Epoch 7/50\n",
            "596/596 [==============================] - 0s 179us/step - loss: 4.2847 - dense_76_loss: 0.5293 - dense_77_loss: 3.3929 - dense_78_loss: 0.3625 - dense_76_f1_m: 0.4263 - dense_76_f1_m_1: 0.4263 - dense_76_f1_m_2: 0.4263 - dense_77_f1_m: 0.8805 - dense_77_f1_m_1: 0.8805 - dense_77_f1_m_2: 0.8805 - dense_78_f1_m: 0.0000e+00 - dense_78_f1_m_1: 0.0000e+00 - dense_78_f1_m_2: 0.0000e+00 - val_loss: 2.8404 - val_dense_76_loss: 0.5392 - val_dense_77_loss: 1.7384 - val_dense_78_loss: 0.5629 - val_dense_76_f1_m: 0.1211 - val_dense_76_f1_m_1: 0.1211 - val_dense_76_f1_m_2: 0.1211 - val_dense_77_f1_m: 0.9696 - val_dense_77_f1_m_1: 0.9696 - val_dense_77_f1_m_2: 0.9696 - val_dense_78_f1_m: 0.0149 - val_dense_78_f1_m_1: 0.0149 - val_dense_78_f1_m_2: 0.0149\n",
            "Epoch 8/50\n",
            "596/596 [==============================] - 0s 158us/step - loss: 4.0451 - dense_76_loss: 0.4729 - dense_77_loss: 3.2206 - dense_78_loss: 0.3517 - dense_76_f1_m: 0.4743 - dense_76_f1_m_1: 0.4743 - dense_76_f1_m_2: 0.4743 - dense_77_f1_m: 0.9084 - dense_77_f1_m_1: 0.9084 - dense_77_f1_m_2: 0.9084 - dense_78_f1_m: 0.0000e+00 - dense_78_f1_m_1: 0.0000e+00 - dense_78_f1_m_2: 0.0000e+00 - val_loss: 2.8043 - val_dense_76_loss: 0.4778 - val_dense_77_loss: 1.7997 - val_dense_78_loss: 0.5268 - val_dense_76_f1_m: 0.5195 - val_dense_76_f1_m_1: 0.5195 - val_dense_76_f1_m_2: 0.5195 - val_dense_77_f1_m: 0.9755 - val_dense_77_f1_m_1: 0.9755 - val_dense_77_f1_m_2: 0.9755 - val_dense_78_f1_m: 0.0073 - val_dense_78_f1_m_1: 0.0073 - val_dense_78_f1_m_2: 0.0073\n",
            "Epoch 9/50\n",
            "596/596 [==============================] - 0s 155us/step - loss: 3.7359 - dense_76_loss: 0.4220 - dense_77_loss: 2.9886 - dense_78_loss: 0.3252 - dense_76_f1_m: 0.5320 - dense_76_f1_m_1: 0.5320 - dense_76_f1_m_2: 0.5320 - dense_77_f1_m: 0.9327 - dense_77_f1_m_1: 0.9327 - dense_77_f1_m_2: 0.9327 - dense_78_f1_m: 0.0000e+00 - dense_78_f1_m_1: 0.0000e+00 - dense_78_f1_m_2: 0.0000e+00 - val_loss: 2.5343 - val_dense_76_loss: 0.4950 - val_dense_77_loss: 1.5064 - val_dense_78_loss: 0.5330 - val_dense_76_f1_m: 0.3320 - val_dense_76_f1_m_1: 0.3320 - val_dense_76_f1_m_2: 0.3320 - val_dense_77_f1_m: 0.9716 - val_dense_77_f1_m_1: 0.9716 - val_dense_77_f1_m_2: 0.9716 - val_dense_78_f1_m: 0.0300 - val_dense_78_f1_m_1: 0.0300 - val_dense_78_f1_m_2: 0.0300\n",
            "Epoch 10/50\n",
            "596/596 [==============================] - 0s 176us/step - loss: 3.5318 - dense_76_loss: 0.4333 - dense_77_loss: 2.7781 - dense_78_loss: 0.3204 - dense_76_f1_m: 0.5493 - dense_76_f1_m_1: 0.5493 - dense_76_f1_m_2: 0.5493 - dense_77_f1_m: 0.9377 - dense_77_f1_m_1: 0.9377 - dense_77_f1_m_2: 0.9377 - dense_78_f1_m: 0.0065 - dense_78_f1_m_1: 0.0065 - dense_78_f1_m_2: 0.0065 - val_loss: 2.3048 - val_dense_76_loss: 0.3682 - val_dense_77_loss: 1.4813 - val_dense_78_loss: 0.4553 - val_dense_76_f1_m: 0.7734 - val_dense_76_f1_m_1: 0.7734 - val_dense_76_f1_m_2: 0.7734 - val_dense_77_f1_m: 0.9774 - val_dense_77_f1_m_1: 0.9774 - val_dense_77_f1_m_2: 0.9774 - val_dense_78_f1_m: 0.0373 - val_dense_78_f1_m_1: 0.0373 - val_dense_78_f1_m_2: 0.0373\n",
            "Epoch 11/50\n",
            "596/596 [==============================] - 0s 192us/step - loss: 3.2071 - dense_76_loss: 0.3766 - dense_77_loss: 2.5333 - dense_78_loss: 0.2973 - dense_76_f1_m: 0.6156 - dense_76_f1_m_1: 0.6156 - dense_76_f1_m_2: 0.6156 - dense_77_f1_m: 0.9448 - dense_77_f1_m_1: 0.9448 - dense_77_f1_m_2: 0.9448 - dense_78_f1_m: 0.0065 - dense_78_f1_m_1: 0.0065 - dense_78_f1_m_2: 0.0065 - val_loss: 2.3118 - val_dense_76_loss: 0.6131 - val_dense_77_loss: 1.0986 - val_dense_78_loss: 0.6001 - val_dense_76_f1_m: 0.0625 - val_dense_76_f1_m_1: 0.0625 - val_dense_76_f1_m_2: 0.0625 - val_dense_77_f1_m: 0.9716 - val_dense_77_f1_m_1: 0.9716 - val_dense_77_f1_m_2: 0.9716 - val_dense_78_f1_m: 0.0298 - val_dense_78_f1_m_1: 0.0298 - val_dense_78_f1_m_2: 0.0298\n",
            "Epoch 12/50\n",
            "596/596 [==============================] - 0s 199us/step - loss: 3.0864 - dense_76_loss: 0.3352 - dense_77_loss: 2.4334 - dense_78_loss: 0.3177 - dense_76_f1_m: 0.5865 - dense_76_f1_m_1: 0.5865 - dense_76_f1_m_2: 0.5865 - dense_77_f1_m: 0.9419 - dense_77_f1_m_1: 0.9419 - dense_77_f1_m_2: 0.9419 - dense_78_f1_m: 0.0189 - dense_78_f1_m_1: 0.0189 - dense_78_f1_m_2: 0.0189 - val_loss: 2.0210 - val_dense_76_loss: 0.4307 - val_dense_77_loss: 1.1202 - val_dense_78_loss: 0.4701 - val_dense_76_f1_m: 0.5195 - val_dense_76_f1_m_1: 0.5195 - val_dense_76_f1_m_2: 0.5195 - val_dense_77_f1_m: 0.9774 - val_dense_77_f1_m_1: 0.9774 - val_dense_77_f1_m_2: 0.9774 - val_dense_78_f1_m: 0.0449 - val_dense_78_f1_m_1: 0.0449 - val_dense_78_f1_m_2: 0.0449\n",
            "Epoch 13/50\n",
            "596/596 [==============================] - 0s 170us/step - loss: 2.7756 - dense_76_loss: 0.3467 - dense_77_loss: 2.1392 - dense_78_loss: 0.2898 - dense_76_f1_m: 0.7340 - dense_76_f1_m_1: 0.7340 - dense_76_f1_m_2: 0.7340 - dense_77_f1_m: 0.9680 - dense_77_f1_m_1: 0.9680 - dense_77_f1_m_2: 0.9680 - dense_78_f1_m: 0.0156 - dense_78_f1_m_1: 0.0156 - dense_78_f1_m_2: 0.0156 - val_loss: 1.7255 - val_dense_76_loss: 0.3752 - val_dense_77_loss: 0.8926 - val_dense_78_loss: 0.4577 - val_dense_76_f1_m: 0.5195 - val_dense_76_f1_m_1: 0.5195 - val_dense_76_f1_m_2: 0.5195 - val_dense_77_f1_m: 0.9794 - val_dense_77_f1_m_1: 0.9794 - val_dense_77_f1_m_2: 0.9794 - val_dense_78_f1_m: 0.2804 - val_dense_78_f1_m_1: 0.2804 - val_dense_78_f1_m_2: 0.2804\n",
            "Epoch 14/50\n",
            "596/596 [==============================] - 0s 167us/step - loss: 2.6314 - dense_76_loss: 0.3261 - dense_77_loss: 2.0046 - dense_78_loss: 0.3007 - dense_76_f1_m: 0.7333 - dense_76_f1_m_1: 0.7333 - dense_76_f1_m_2: 0.7333 - dense_77_f1_m: 0.9633 - dense_77_f1_m_1: 0.9633 - dense_77_f1_m_2: 0.9633 - dense_78_f1_m: 0.0251 - dense_78_f1_m_1: 0.0251 - dense_78_f1_m_2: 0.0251 - val_loss: 1.5508 - val_dense_76_loss: 0.3116 - val_dense_77_loss: 0.8111 - val_dense_78_loss: 0.4281 - val_dense_76_f1_m: 0.8125 - val_dense_76_f1_m_1: 0.8125 - val_dense_76_f1_m_2: 0.8125 - val_dense_77_f1_m: 0.9772 - val_dense_77_f1_m_1: 0.9772 - val_dense_77_f1_m_2: 0.9772 - val_dense_78_f1_m: 0.5476 - val_dense_78_f1_m_1: 0.5476 - val_dense_78_f1_m_2: 0.5476\n",
            "Epoch 15/50\n",
            "596/596 [==============================] - 0s 173us/step - loss: 2.4086 - dense_76_loss: 0.2868 - dense_77_loss: 1.8473 - dense_78_loss: 0.2746 - dense_76_f1_m: 0.7130 - dense_76_f1_m_1: 0.7130 - dense_76_f1_m_2: 0.7130 - dense_77_f1_m: 0.9658 - dense_77_f1_m_1: 0.9658 - dense_77_f1_m_2: 0.9658 - dense_78_f1_m: 0.0924 - dense_78_f1_m_1: 0.0924 - dense_78_f1_m_2: 0.0924 - val_loss: 1.4412 - val_dense_76_loss: 0.3103 - val_dense_77_loss: 0.7289 - val_dense_78_loss: 0.4020 - val_dense_76_f1_m: 0.7109 - val_dense_76_f1_m_1: 0.7109 - val_dense_76_f1_m_2: 0.7109 - val_dense_77_f1_m: 0.9813 - val_dense_77_f1_m_1: 0.9813 - val_dense_77_f1_m_2: 0.9813 - val_dense_78_f1_m: 0.6787 - val_dense_78_f1_m_1: 0.6787 - val_dense_78_f1_m_2: 0.6787\n",
            "Epoch 16/50\n",
            "596/596 [==============================] - 0s 168us/step - loss: 2.2135 - dense_76_loss: 0.2811 - dense_77_loss: 1.6595 - dense_78_loss: 0.2728 - dense_76_f1_m: 0.7697 - dense_76_f1_m_1: 0.7697 - dense_76_f1_m_2: 0.7697 - dense_77_f1_m: 0.9749 - dense_77_f1_m_1: 0.9749 - dense_77_f1_m_2: 0.9749 - dense_78_f1_m: 0.0894 - dense_78_f1_m_1: 0.0894 - dense_78_f1_m_2: 0.0894 - val_loss: 1.5713 - val_dense_76_loss: 0.4441 - val_dense_77_loss: 0.6430 - val_dense_78_loss: 0.4842 - val_dense_76_f1_m: 0.4727 - val_dense_76_f1_m_1: 0.4727 - val_dense_76_f1_m_2: 0.4727 - val_dense_77_f1_m: 0.9754 - val_dense_77_f1_m_1: 0.9754 - val_dense_77_f1_m_2: 0.9754 - val_dense_78_f1_m: 0.7014 - val_dense_78_f1_m_1: 0.7014 - val_dense_78_f1_m_2: 0.7014\n",
            "Epoch 17/50\n",
            "596/596 [==============================] - 0s 169us/step - loss: 2.1241 - dense_76_loss: 0.2758 - dense_77_loss: 1.5758 - dense_78_loss: 0.2725 - dense_76_f1_m: 0.6942 - dense_76_f1_m_1: 0.6942 - dense_76_f1_m_2: 0.6942 - dense_77_f1_m: 0.9690 - dense_77_f1_m_1: 0.9690 - dense_77_f1_m_2: 0.9690 - dense_78_f1_m: 0.2286 - dense_78_f1_m_1: 0.2286 - dense_78_f1_m_2: 0.2286 - val_loss: 1.4743 - val_dense_76_loss: 0.3445 - val_dense_77_loss: 0.6905 - val_dense_78_loss: 0.4393 - val_dense_76_f1_m: 0.5391 - val_dense_76_f1_m_1: 0.5391 - val_dense_76_f1_m_2: 0.5391 - val_dense_77_f1_m: 0.9794 - val_dense_77_f1_m_1: 0.9794 - val_dense_77_f1_m_2: 0.9794 - val_dense_78_f1_m: 0.6059 - val_dense_78_f1_m_1: 0.6059 - val_dense_78_f1_m_2: 0.6059\n",
            "Epoch 18/50\n",
            "596/596 [==============================] - 0s 196us/step - loss: 1.8449 - dense_76_loss: 0.2842 - dense_77_loss: 1.2725 - dense_78_loss: 0.2882 - dense_76_f1_m: 0.7994 - dense_76_f1_m_1: 0.7994 - dense_76_f1_m_2: 0.7994 - dense_77_f1_m: 0.9806 - dense_77_f1_m_1: 0.9806 - dense_77_f1_m_2: 0.9806 - dense_78_f1_m: 0.2112 - dense_78_f1_m_1: 0.2112 - dense_78_f1_m_2: 0.2112 - val_loss: 1.4521 - val_dense_76_loss: 0.4010 - val_dense_77_loss: 0.6012 - val_dense_78_loss: 0.4499 - val_dense_76_f1_m: 0.5391 - val_dense_76_f1_m_1: 0.5391 - val_dense_76_f1_m_2: 0.5391 - val_dense_77_f1_m: 0.9795 - val_dense_77_f1_m_1: 0.9795 - val_dense_77_f1_m_2: 0.9795 - val_dense_78_f1_m: 0.8380 - val_dense_78_f1_m_1: 0.8380 - val_dense_78_f1_m_2: 0.8380\n",
            "Epoch 19/50\n",
            "596/596 [==============================] - 0s 162us/step - loss: 1.8506 - dense_76_loss: 0.2619 - dense_77_loss: 1.3051 - dense_78_loss: 0.2836 - dense_76_f1_m: 0.7412 - dense_76_f1_m_1: 0.7412 - dense_76_f1_m_2: 0.7412 - dense_77_f1_m: 0.9739 - dense_77_f1_m_1: 0.9739 - dense_77_f1_m_2: 0.9739 - dense_78_f1_m: 0.3995 - dense_78_f1_m_1: 0.3995 - dense_78_f1_m_2: 0.3995 - val_loss: 1.4628 - val_dense_76_loss: 0.3672 - val_dense_77_loss: 0.6801 - val_dense_78_loss: 0.4156 - val_dense_76_f1_m: 0.5391 - val_dense_76_f1_m_1: 0.5391 - val_dense_76_f1_m_2: 0.5391 - val_dense_77_f1_m: 0.9793 - val_dense_77_f1_m_1: 0.9793 - val_dense_77_f1_m_2: 0.9793 - val_dense_78_f1_m: 0.9126 - val_dense_78_f1_m_1: 0.9126 - val_dense_78_f1_m_2: 0.9126\n",
            "Epoch 20/50\n",
            "596/596 [==============================] - 0s 171us/step - loss: 1.6804 - dense_76_loss: 0.2643 - dense_77_loss: 1.1488 - dense_78_loss: 0.2673 - dense_76_f1_m: 0.8189 - dense_76_f1_m_1: 0.8189 - dense_76_f1_m_2: 0.8189 - dense_77_f1_m: 0.9750 - dense_77_f1_m_1: 0.9750 - dense_77_f1_m_2: 0.9750 - dense_78_f1_m: 0.4388 - dense_78_f1_m_1: 0.4388 - dense_78_f1_m_2: 0.4388 - val_loss: 1.2323 - val_dense_76_loss: 0.2762 - val_dense_77_loss: 0.5925 - val_dense_78_loss: 0.3636 - val_dense_76_f1_m: 0.7500 - val_dense_76_f1_m_1: 0.7500 - val_dense_76_f1_m_2: 0.7500 - val_dense_77_f1_m: 0.9793 - val_dense_77_f1_m_1: 0.9793 - val_dense_77_f1_m_2: 0.9793 - val_dense_78_f1_m: 0.9299 - val_dense_78_f1_m_1: 0.9299 - val_dense_78_f1_m_2: 0.9299\n",
            "Epoch 21/50\n",
            "596/596 [==============================] - 0s 159us/step - loss: 1.6912 - dense_76_loss: 0.2592 - dense_77_loss: 1.1559 - dense_78_loss: 0.2761 - dense_76_f1_m: 0.6861 - dense_76_f1_m_1: 0.6861 - dense_76_f1_m_2: 0.6861 - dense_77_f1_m: 0.9773 - dense_77_f1_m_1: 0.9773 - dense_77_f1_m_2: 0.9773 - dense_78_f1_m: 0.4934 - dense_78_f1_m_1: 0.4934 - dense_78_f1_m_2: 0.4934 - val_loss: 1.2261 - val_dense_76_loss: 0.3136 - val_dense_77_loss: 0.5258 - val_dense_78_loss: 0.3867 - val_dense_76_f1_m: 0.6328 - val_dense_76_f1_m_1: 0.6328 - val_dense_76_f1_m_2: 0.6328 - val_dense_77_f1_m: 0.9794 - val_dense_77_f1_m_1: 0.9794 - val_dense_77_f1_m_2: 0.9794 - val_dense_78_f1_m: 0.9464 - val_dense_78_f1_m_1: 0.9464 - val_dense_78_f1_m_2: 0.9464\n",
            "Epoch 22/50\n",
            "596/596 [==============================] - 0s 187us/step - loss: 1.4443 - dense_76_loss: 0.2063 - dense_77_loss: 0.9802 - dense_78_loss: 0.2577 - dense_76_f1_m: 0.8456 - dense_76_f1_m_1: 0.8456 - dense_76_f1_m_2: 0.8456 - dense_77_f1_m: 0.9795 - dense_77_f1_m_1: 0.9795 - dense_77_f1_m_2: 0.9795 - dense_78_f1_m: 0.5968 - dense_78_f1_m_1: 0.5968 - dense_78_f1_m_2: 0.5968 - val_loss: 1.0976 - val_dense_76_loss: 0.1901 - val_dense_77_loss: 0.5903 - val_dense_78_loss: 0.3171 - val_dense_76_f1_m: 0.9208 - val_dense_76_f1_m_1: 0.9208 - val_dense_76_f1_m_2: 0.9208 - val_dense_77_f1_m: 0.9751 - val_dense_77_f1_m_1: 0.9751 - val_dense_77_f1_m_2: 0.9751 - val_dense_78_f1_m: 0.9252 - val_dense_78_f1_m_1: 0.9252 - val_dense_78_f1_m_2: 0.9252\n",
            "Epoch 23/50\n",
            "596/596 [==============================] - 0s 178us/step - loss: 1.4339 - dense_76_loss: 0.2037 - dense_77_loss: 0.9542 - dense_78_loss: 0.2760 - dense_76_f1_m: 0.8880 - dense_76_f1_m_1: 0.8880 - dense_76_f1_m_2: 0.8880 - dense_77_f1_m: 0.9775 - dense_77_f1_m_1: 0.9775 - dense_77_f1_m_2: 0.9775 - dense_78_f1_m: 0.6224 - dense_78_f1_m_1: 0.6224 - dense_78_f1_m_2: 0.6224 - val_loss: 1.1064 - val_dense_76_loss: 0.2104 - val_dense_77_loss: 0.5476 - val_dense_78_loss: 0.3483 - val_dense_76_f1_m: 0.7969 - val_dense_76_f1_m_1: 0.7969 - val_dense_76_f1_m_2: 0.7969 - val_dense_77_f1_m: 0.9774 - val_dense_77_f1_m_1: 0.9774 - val_dense_77_f1_m_2: 0.9774 - val_dense_78_f1_m: 0.9310 - val_dense_78_f1_m_1: 0.9310 - val_dense_78_f1_m_2: 0.9310\n",
            "Epoch 24/50\n",
            "596/596 [==============================] - 0s 171us/step - loss: 1.4227 - dense_76_loss: 0.2705 - dense_77_loss: 0.8193 - dense_78_loss: 0.3329 - dense_76_f1_m: 0.7457 - dense_76_f1_m_1: 0.7457 - dense_76_f1_m_2: 0.7457 - dense_77_f1_m: 0.9764 - dense_77_f1_m_1: 0.9764 - dense_77_f1_m_2: 0.9764 - dense_78_f1_m: 0.6576 - dense_78_f1_m_1: 0.6576 - dense_78_f1_m_2: 0.6576 - val_loss: 1.0550 - val_dense_76_loss: 0.1361 - val_dense_77_loss: 0.6043 - val_dense_78_loss: 0.3146 - val_dense_76_f1_m: 0.8873 - val_dense_76_f1_m_1: 0.8873 - val_dense_76_f1_m_2: 0.8873 - val_dense_77_f1_m: 0.9793 - val_dense_77_f1_m_1: 0.9793 - val_dense_77_f1_m_2: 0.9793 - val_dense_78_f1_m: 0.8825 - val_dense_78_f1_m_1: 0.8825 - val_dense_78_f1_m_2: 0.8825\n",
            "Epoch 25/50\n",
            "596/596 [==============================] - 0s 183us/step - loss: 1.2880 - dense_76_loss: 0.2280 - dense_77_loss: 0.7978 - dense_78_loss: 0.2622 - dense_76_f1_m: 0.8217 - dense_76_f1_m_1: 0.8217 - dense_76_f1_m_2: 0.8217 - dense_77_f1_m: 0.9752 - dense_77_f1_m_1: 0.9752 - dense_77_f1_m_2: 0.9752 - dense_78_f1_m: 0.7166 - dense_78_f1_m_1: 0.7166 - dense_78_f1_m_2: 0.7166 - val_loss: 0.9792 - val_dense_76_loss: 0.1371 - val_dense_77_loss: 0.5482 - val_dense_78_loss: 0.2939 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9772 - val_dense_77_f1_m_1: 0.9772 - val_dense_77_f1_m_2: 0.9772 - val_dense_78_f1_m: 0.9382 - val_dense_78_f1_m_1: 0.9382 - val_dense_78_f1_m_2: 0.9382\n",
            "Epoch 26/50\n",
            "596/596 [==============================] - 0s 169us/step - loss: 1.1854 - dense_76_loss: 0.2144 - dense_77_loss: 0.7163 - dense_78_loss: 0.2547 - dense_76_f1_m: 0.7991 - dense_76_f1_m_1: 0.7991 - dense_76_f1_m_2: 0.7991 - dense_77_f1_m: 0.9720 - dense_77_f1_m_1: 0.9720 - dense_77_f1_m_2: 0.9720 - dense_78_f1_m: 0.7331 - dense_78_f1_m_1: 0.7331 - dense_78_f1_m_2: 0.7331 - val_loss: 1.0555 - val_dense_76_loss: 0.1086 - val_dense_77_loss: 0.6797 - val_dense_78_loss: 0.2672 - val_dense_76_f1_m: 0.9040 - val_dense_76_f1_m_1: 0.9040 - val_dense_76_f1_m_2: 0.9040 - val_dense_77_f1_m: 0.9832 - val_dense_77_f1_m_1: 0.9832 - val_dense_77_f1_m_2: 0.9832 - val_dense_78_f1_m: 0.9120 - val_dense_78_f1_m_1: 0.9120 - val_dense_78_f1_m_2: 0.9120\n",
            "Epoch 27/50\n",
            "596/596 [==============================] - 0s 176us/step - loss: 1.1350 - dense_76_loss: 0.1782 - dense_77_loss: 0.7218 - dense_78_loss: 0.2350 - dense_76_f1_m: 0.7828 - dense_76_f1_m_1: 0.7828 - dense_76_f1_m_2: 0.7828 - dense_77_f1_m: 0.9804 - dense_77_f1_m_1: 0.9804 - dense_77_f1_m_2: 0.9804 - dense_78_f1_m: 0.7874 - dense_78_f1_m_1: 0.7874 - dense_78_f1_m_2: 0.7874 - val_loss: 0.9808 - val_dense_76_loss: 0.1535 - val_dense_77_loss: 0.5273 - val_dense_78_loss: 0.3000 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9772 - val_dense_77_f1_m_1: 0.9772 - val_dense_77_f1_m_2: 0.9772 - val_dense_78_f1_m: 0.9426 - val_dense_78_f1_m_1: 0.9426 - val_dense_78_f1_m_2: 0.9426\n",
            "Epoch 28/50\n",
            "596/596 [==============================] - 0s 176us/step - loss: 1.1393 - dense_76_loss: 0.2287 - dense_77_loss: 0.6345 - dense_78_loss: 0.2762 - dense_76_f1_m: 0.7568 - dense_76_f1_m_1: 0.7568 - dense_76_f1_m_2: 0.7568 - dense_77_f1_m: 0.9675 - dense_77_f1_m_1: 0.9675 - dense_77_f1_m_2: 0.9675 - dense_78_f1_m: 0.7851 - dense_78_f1_m_1: 0.7851 - dense_78_f1_m_2: 0.7851 - val_loss: 0.9300 - val_dense_76_loss: 0.1638 - val_dense_77_loss: 0.4538 - val_dense_78_loss: 0.3125 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9814 - val_dense_77_f1_m_1: 0.9814 - val_dense_77_f1_m_2: 0.9814 - val_dense_78_f1_m: 0.9422 - val_dense_78_f1_m_1: 0.9422 - val_dense_78_f1_m_2: 0.9422\n",
            "Epoch 29/50\n",
            "596/596 [==============================] - 0s 177us/step - loss: 1.0553 - dense_76_loss: 0.1928 - dense_77_loss: 0.5904 - dense_78_loss: 0.2721 - dense_76_f1_m: 0.8480 - dense_76_f1_m_1: 0.8480 - dense_76_f1_m_2: 0.8480 - dense_77_f1_m: 0.9714 - dense_77_f1_m_1: 0.9714 - dense_77_f1_m_2: 0.9714 - dense_78_f1_m: 0.8068 - dense_78_f1_m_1: 0.8068 - dense_78_f1_m_2: 0.8068 - val_loss: 1.0563 - val_dense_76_loss: 0.2936 - val_dense_77_loss: 0.3925 - val_dense_78_loss: 0.3701 - val_dense_76_f1_m: 0.6406 - val_dense_76_f1_m_1: 0.6406 - val_dense_76_f1_m_2: 0.6406 - val_dense_77_f1_m: 0.9772 - val_dense_77_f1_m_1: 0.9772 - val_dense_77_f1_m_2: 0.9772 - val_dense_78_f1_m: 0.9472 - val_dense_78_f1_m_1: 0.9472 - val_dense_78_f1_m_2: 0.9472\n",
            "Epoch 30/50\n",
            "596/596 [==============================] - 0s 176us/step - loss: 1.0216 - dense_76_loss: 0.1845 - dense_77_loss: 0.5768 - dense_78_loss: 0.2603 - dense_76_f1_m: 0.8323 - dense_76_f1_m_1: 0.8323 - dense_76_f1_m_2: 0.8323 - dense_77_f1_m: 0.9743 - dense_77_f1_m_1: 0.9743 - dense_77_f1_m_2: 0.9743 - dense_78_f1_m: 0.8154 - dense_78_f1_m_1: 0.8154 - dense_78_f1_m_2: 0.8154 - val_loss: 0.8812 - val_dense_76_loss: 0.1822 - val_dense_77_loss: 0.4071 - val_dense_78_loss: 0.2919 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9814 - val_dense_77_f1_m_1: 0.9814 - val_dense_77_f1_m_2: 0.9814 - val_dense_78_f1_m: 0.9394 - val_dense_78_f1_m_1: 0.9394 - val_dense_78_f1_m_2: 0.9394\n",
            "Epoch 31/50\n",
            "596/596 [==============================] - 0s 166us/step - loss: 0.8942 - dense_76_loss: 0.1682 - dense_77_loss: 0.4956 - dense_78_loss: 0.2303 - dense_76_f1_m: 0.8702 - dense_76_f1_m_1: 0.8702 - dense_76_f1_m_2: 0.8702 - dense_77_f1_m: 0.9804 - dense_77_f1_m_1: 0.9804 - dense_77_f1_m_2: 0.9804 - dense_78_f1_m: 0.8380 - dense_78_f1_m_1: 0.8380 - dense_78_f1_m_2: 0.8380 - val_loss: 0.8490 - val_dense_76_loss: 0.1266 - val_dense_77_loss: 0.4655 - val_dense_78_loss: 0.2568 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9794 - val_dense_77_f1_m_1: 0.9794 - val_dense_77_f1_m_2: 0.9794 - val_dense_78_f1_m: 0.9316 - val_dense_78_f1_m_1: 0.9316 - val_dense_78_f1_m_2: 0.9316\n",
            "Epoch 32/50\n",
            "596/596 [==============================] - 0s 179us/step - loss: 0.9234 - dense_76_loss: 0.1634 - dense_77_loss: 0.5250 - dense_78_loss: 0.2350 - dense_76_f1_m: 0.7769 - dense_76_f1_m_1: 0.7769 - dense_76_f1_m_2: 0.7769 - dense_77_f1_m: 0.9723 - dense_77_f1_m_1: 0.9723 - dense_77_f1_m_2: 0.9723 - dense_78_f1_m: 0.8314 - dense_78_f1_m_1: 0.8314 - dense_78_f1_m_2: 0.8314 - val_loss: 0.8321 - val_dense_76_loss: 0.0939 - val_dense_77_loss: 0.4846 - val_dense_78_loss: 0.2535 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9833 - val_dense_77_f1_m_1: 0.9833 - val_dense_77_f1_m_2: 0.9833 - val_dense_78_f1_m: 0.9196 - val_dense_78_f1_m_1: 0.9196 - val_dense_78_f1_m_2: 0.9196\n",
            "Epoch 33/50\n",
            "596/596 [==============================] - 0s 171us/step - loss: 0.8692 - dense_76_loss: 0.1383 - dense_77_loss: 0.4804 - dense_78_loss: 0.2505 - dense_76_f1_m: 0.8782 - dense_76_f1_m_1: 0.8782 - dense_76_f1_m_2: 0.8782 - dense_77_f1_m: 0.9698 - dense_77_f1_m_1: 0.9698 - dense_77_f1_m_2: 0.9698 - dense_78_f1_m: 0.8491 - dense_78_f1_m_1: 0.8491 - dense_78_f1_m_2: 0.8491 - val_loss: 0.9786 - val_dense_76_loss: 0.2962 - val_dense_77_loss: 0.3298 - val_dense_78_loss: 0.3526 - val_dense_76_f1_m: 0.7109 - val_dense_76_f1_m_1: 0.7109 - val_dense_76_f1_m_2: 0.7109 - val_dense_77_f1_m: 0.9734 - val_dense_77_f1_m_1: 0.9734 - val_dense_77_f1_m_2: 0.9734 - val_dense_78_f1_m: 0.9495 - val_dense_78_f1_m_1: 0.9495 - val_dense_78_f1_m_2: 0.9495\n",
            "Epoch 34/50\n",
            "596/596 [==============================] - 0s 169us/step - loss: 0.9256 - dense_76_loss: 0.1941 - dense_77_loss: 0.4695 - dense_78_loss: 0.2619 - dense_76_f1_m: 0.7780 - dense_76_f1_m_1: 0.7780 - dense_76_f1_m_2: 0.7780 - dense_77_f1_m: 0.9669 - dense_77_f1_m_1: 0.9669 - dense_77_f1_m_2: 0.9669 - dense_78_f1_m: 0.8520 - dense_78_f1_m_1: 0.8520 - dense_78_f1_m_2: 0.8520 - val_loss: 0.7493 - val_dense_76_loss: 0.1052 - val_dense_77_loss: 0.3835 - val_dense_78_loss: 0.2606 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9752 - val_dense_77_f1_m_1: 0.9752 - val_dense_77_f1_m_2: 0.9752 - val_dense_78_f1_m: 0.9294 - val_dense_78_f1_m_1: 0.9294 - val_dense_78_f1_m_2: 0.9294\n",
            "Epoch 35/50\n",
            "596/596 [==============================] - 0s 195us/step - loss: 0.8027 - dense_76_loss: 0.1563 - dense_77_loss: 0.3867 - dense_78_loss: 0.2596 - dense_76_f1_m: 0.8100 - dense_76_f1_m_1: 0.8100 - dense_76_f1_m_2: 0.8100 - dense_77_f1_m: 0.9749 - dense_77_f1_m_1: 0.9749 - dense_77_f1_m_2: 0.9749 - dense_78_f1_m: 0.8623 - dense_78_f1_m_1: 0.8623 - dense_78_f1_m_2: 0.8623 - val_loss: 0.7383 - val_dense_76_loss: 0.1013 - val_dense_77_loss: 0.3889 - val_dense_78_loss: 0.2481 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9793 - val_dense_77_f1_m_1: 0.9793 - val_dense_77_f1_m_2: 0.9793 - val_dense_78_f1_m: 0.9297 - val_dense_78_f1_m_1: 0.9297 - val_dense_78_f1_m_2: 0.9297\n",
            "Epoch 36/50\n",
            "596/596 [==============================] - 0s 181us/step - loss: 0.8614 - dense_76_loss: 0.1952 - dense_77_loss: 0.4270 - dense_78_loss: 0.2391 - dense_76_f1_m: 0.7836 - dense_76_f1_m_1: 0.7836 - dense_76_f1_m_2: 0.7836 - dense_77_f1_m: 0.9599 - dense_77_f1_m_1: 0.9599 - dense_77_f1_m_2: 0.9599 - dense_78_f1_m: 0.8578 - dense_78_f1_m_1: 0.8578 - dense_78_f1_m_2: 0.8578 - val_loss: 0.6926 - val_dense_76_loss: 0.1060 - val_dense_77_loss: 0.3303 - val_dense_78_loss: 0.2563 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9772 - val_dense_77_f1_m_1: 0.9772 - val_dense_77_f1_m_2: 0.9772 - val_dense_78_f1_m: 0.9378 - val_dense_78_f1_m_1: 0.9378 - val_dense_78_f1_m_2: 0.9378\n",
            "Epoch 37/50\n",
            "596/596 [==============================] - 0s 176us/step - loss: 0.6836 - dense_76_loss: 0.1283 - dense_77_loss: 0.3058 - dense_78_loss: 0.2495 - dense_76_f1_m: 0.8437 - dense_76_f1_m_1: 0.8437 - dense_76_f1_m_2: 0.8437 - dense_77_f1_m: 0.9816 - dense_77_f1_m_1: 0.9816 - dense_77_f1_m_2: 0.9816 - dense_78_f1_m: 0.8657 - dense_78_f1_m_1: 0.8657 - dense_78_f1_m_2: 0.8657 - val_loss: 0.6800 - val_dense_76_loss: 0.0972 - val_dense_77_loss: 0.3366 - val_dense_78_loss: 0.2462 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9833 - val_dense_77_f1_m_1: 0.9833 - val_dense_77_f1_m_2: 0.9833 - val_dense_78_f1_m: 0.9245 - val_dense_78_f1_m_1: 0.9245 - val_dense_78_f1_m_2: 0.9245\n",
            "Epoch 38/50\n",
            "596/596 [==============================] - 0s 166us/step - loss: 0.9544 - dense_76_loss: 0.2006 - dense_77_loss: 0.4574 - dense_78_loss: 0.2964 - dense_76_f1_m: 0.7518 - dense_76_f1_m_1: 0.7518 - dense_76_f1_m_2: 0.7518 - dense_77_f1_m: 0.9655 - dense_77_f1_m_1: 0.9655 - dense_77_f1_m_2: 0.9655 - dense_78_f1_m: 0.8472 - dense_78_f1_m_1: 0.8472 - dense_78_f1_m_2: 0.8472 - val_loss: 1.1081 - val_dense_76_loss: 0.4210 - val_dense_77_loss: 0.2462 - val_dense_78_loss: 0.4409 - val_dense_76_f1_m: 0.5391 - val_dense_76_f1_m_1: 0.5391 - val_dense_76_f1_m_2: 0.5391 - val_dense_77_f1_m: 0.9695 - val_dense_77_f1_m_1: 0.9695 - val_dense_77_f1_m_2: 0.9695 - val_dense_78_f1_m: 0.9378 - val_dense_78_f1_m_1: 0.9378 - val_dense_78_f1_m_2: 0.9378\n",
            "Epoch 39/50\n",
            "596/596 [==============================] - 0s 164us/step - loss: 0.8438 - dense_76_loss: 0.1533 - dense_77_loss: 0.4620 - dense_78_loss: 0.2284 - dense_76_f1_m: 0.8013 - dense_76_f1_m_1: 0.8013 - dense_76_f1_m_2: 0.8013 - dense_77_f1_m: 0.9635 - dense_77_f1_m_1: 0.9635 - dense_77_f1_m_2: 0.9635 - dense_78_f1_m: 0.8558 - dense_78_f1_m_1: 0.8558 - dense_78_f1_m_2: 0.8558 - val_loss: 0.6703 - val_dense_76_loss: 0.0958 - val_dense_77_loss: 0.3109 - val_dense_78_loss: 0.2637 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9813 - val_dense_77_f1_m_1: 0.9813 - val_dense_77_f1_m_2: 0.9813 - val_dense_78_f1_m: 0.9298 - val_dense_78_f1_m_1: 0.9298 - val_dense_78_f1_m_2: 0.9298\n",
            "Epoch 40/50\n",
            "596/596 [==============================] - 0s 186us/step - loss: 0.7777 - dense_76_loss: 0.1380 - dense_77_loss: 0.3708 - dense_78_loss: 0.2690 - dense_76_f1_m: 0.8164 - dense_76_f1_m_1: 0.8164 - dense_76_f1_m_2: 0.8164 - dense_77_f1_m: 0.9575 - dense_77_f1_m_1: 0.9575 - dense_77_f1_m_2: 0.9575 - dense_78_f1_m: 0.8667 - dense_78_f1_m_1: 0.8667 - dense_78_f1_m_2: 0.8667 - val_loss: 0.7768 - val_dense_76_loss: 0.2071 - val_dense_77_loss: 0.2519 - val_dense_78_loss: 0.3178 - val_dense_76_f1_m: 0.8750 - val_dense_76_f1_m_1: 0.8750 - val_dense_76_f1_m_2: 0.8750 - val_dense_77_f1_m: 0.9714 - val_dense_77_f1_m_1: 0.9714 - val_dense_77_f1_m_2: 0.9714 - val_dense_78_f1_m: 0.9515 - val_dense_78_f1_m_1: 0.9515 - val_dense_78_f1_m_2: 0.9515\n",
            "Epoch 41/50\n",
            "596/596 [==============================] - 0s 165us/step - loss: 0.7541 - dense_76_loss: 0.1480 - dense_77_loss: 0.3525 - dense_78_loss: 0.2535 - dense_76_f1_m: 0.7868 - dense_76_f1_m_1: 0.7868 - dense_76_f1_m_2: 0.7868 - dense_77_f1_m: 0.9630 - dense_77_f1_m_1: 0.9630 - dense_77_f1_m_2: 0.9630 - dense_78_f1_m: 0.8691 - dense_78_f1_m_1: 0.8691 - dense_78_f1_m_2: 0.8691 - val_loss: 0.7069 - val_dense_76_loss: 0.1731 - val_dense_77_loss: 0.2408 - val_dense_78_loss: 0.2930 - val_dense_76_f1_m: 0.8555 - val_dense_76_f1_m_1: 0.8555 - val_dense_76_f1_m_2: 0.8555 - val_dense_77_f1_m: 0.9773 - val_dense_77_f1_m_1: 0.9773 - val_dense_77_f1_m_2: 0.9773 - val_dense_78_f1_m: 0.9418 - val_dense_78_f1_m_1: 0.9418 - val_dense_78_f1_m_2: 0.9418\n",
            "Epoch 42/50\n",
            "596/596 [==============================] - 0s 183us/step - loss: 0.7689 - dense_76_loss: 0.1295 - dense_77_loss: 0.4000 - dense_78_loss: 0.2394 - dense_76_f1_m: 0.7771 - dense_76_f1_m_1: 0.7771 - dense_76_f1_m_2: 0.7771 - dense_77_f1_m: 0.9581 - dense_77_f1_m_1: 0.9581 - dense_77_f1_m_2: 0.9581 - dense_78_f1_m: 0.8651 - dense_78_f1_m_1: 0.8651 - dense_78_f1_m_2: 0.8651 - val_loss: 0.6536 - val_dense_76_loss: 0.1386 - val_dense_77_loss: 0.2493 - val_dense_78_loss: 0.2657 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9774 - val_dense_77_f1_m_1: 0.9774 - val_dense_77_f1_m_2: 0.9774 - val_dense_78_f1_m: 0.9459 - val_dense_78_f1_m_1: 0.9459 - val_dense_78_f1_m_2: 0.9459\n",
            "Epoch 43/50\n",
            "596/596 [==============================] - 0s 161us/step - loss: 0.6924 - dense_76_loss: 0.1341 - dense_77_loss: 0.2903 - dense_78_loss: 0.2680 - dense_76_f1_m: 0.8575 - dense_76_f1_m_1: 0.8575 - dense_76_f1_m_2: 0.8575 - dense_77_f1_m: 0.9657 - dense_77_f1_m_1: 0.9657 - dense_77_f1_m_2: 0.9657 - dense_78_f1_m: 0.8739 - dense_78_f1_m_1: 0.8739 - dense_78_f1_m_2: 0.8739 - val_loss: 0.5668 - val_dense_76_loss: 0.0958 - val_dense_77_loss: 0.2333 - val_dense_78_loss: 0.2377 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9814 - val_dense_77_f1_m_1: 0.9814 - val_dense_77_f1_m_2: 0.9814 - val_dense_78_f1_m: 0.9425 - val_dense_78_f1_m_1: 0.9425 - val_dense_78_f1_m_2: 0.9425\n",
            "Epoch 44/50\n",
            "596/596 [==============================] - 0s 175us/step - loss: 0.6467 - dense_76_loss: 0.1161 - dense_77_loss: 0.3123 - dense_78_loss: 0.2183 - dense_76_f1_m: 0.7985 - dense_76_f1_m_1: 0.7985 - dense_76_f1_m_2: 0.7985 - dense_77_f1_m: 0.9604 - dense_77_f1_m_1: 0.9604 - dense_77_f1_m_2: 0.9604 - dense_78_f1_m: 0.8708 - dense_78_f1_m_1: 0.8708 - dense_78_f1_m_2: 0.8708 - val_loss: 0.6057 - val_dense_76_loss: 0.1355 - val_dense_77_loss: 0.2224 - val_dense_78_loss: 0.2479 - val_dense_76_f1_m: 0.9219 - val_dense_76_f1_m_1: 0.9219 - val_dense_76_f1_m_2: 0.9219 - val_dense_77_f1_m: 0.9734 - val_dense_77_f1_m_1: 0.9734 - val_dense_77_f1_m_2: 0.9734 - val_dense_78_f1_m: 0.9440 - val_dense_78_f1_m_1: 0.9440 - val_dense_78_f1_m_2: 0.9440\n",
            "Epoch 45/50\n",
            "596/596 [==============================] - 0s 185us/step - loss: 0.6344 - dense_76_loss: 0.1320 - dense_77_loss: 0.2842 - dense_78_loss: 0.2182 - dense_76_f1_m: 0.8744 - dense_76_f1_m_1: 0.8744 - dense_76_f1_m_2: 0.8744 - dense_77_f1_m: 0.9630 - dense_77_f1_m_1: 0.9630 - dense_77_f1_m_2: 0.9630 - dense_78_f1_m: 0.8806 - dense_78_f1_m_1: 0.8806 - dense_78_f1_m_2: 0.8806 - val_loss: 0.5352 - val_dense_76_loss: 0.0549 - val_dense_77_loss: 0.2512 - val_dense_78_loss: 0.2291 - val_dense_76_f1_m: 0.9364 - val_dense_76_f1_m_1: 0.9364 - val_dense_76_f1_m_2: 0.9364 - val_dense_77_f1_m: 0.9733 - val_dense_77_f1_m_1: 0.9733 - val_dense_77_f1_m_2: 0.9733 - val_dense_78_f1_m: 0.9244 - val_dense_78_f1_m_1: 0.9244 - val_dense_78_f1_m_2: 0.9244\n",
            "Epoch 46/50\n",
            "596/596 [==============================] - 0s 165us/step - loss: 0.6727 - dense_76_loss: 0.1390 - dense_77_loss: 0.3009 - dense_78_loss: 0.2328 - dense_76_f1_m: 0.7757 - dense_76_f1_m_1: 0.7757 - dense_76_f1_m_2: 0.7757 - dense_77_f1_m: 0.9639 - dense_77_f1_m_1: 0.9639 - dense_77_f1_m_2: 0.9639 - dense_78_f1_m: 0.8584 - dense_78_f1_m_1: 0.8584 - dense_78_f1_m_2: 0.8584 - val_loss: 0.5195 - val_dense_76_loss: 0.0598 - val_dense_77_loss: 0.2307 - val_dense_78_loss: 0.2291 - val_dense_76_f1_m: 0.9442 - val_dense_76_f1_m_1: 0.9442 - val_dense_76_f1_m_2: 0.9442 - val_dense_77_f1_m: 0.9775 - val_dense_77_f1_m_1: 0.9775 - val_dense_77_f1_m_2: 0.9775 - val_dense_78_f1_m: 0.9299 - val_dense_78_f1_m_1: 0.9299 - val_dense_78_f1_m_2: 0.9299\n",
            "Epoch 47/50\n",
            "596/596 [==============================] - 0s 171us/step - loss: 0.7491 - dense_76_loss: 0.1550 - dense_77_loss: 0.3041 - dense_78_loss: 0.2900 - dense_76_f1_m: 0.8396 - dense_76_f1_m_1: 0.8396 - dense_76_f1_m_2: 0.8396 - dense_77_f1_m: 0.9671 - dense_77_f1_m_1: 0.9671 - dense_77_f1_m_2: 0.9671 - dense_78_f1_m: 0.8772 - dense_78_f1_m_1: 0.8772 - dense_78_f1_m_2: 0.8772 - val_loss: 0.7312 - val_dense_76_loss: 0.2265 - val_dense_77_loss: 0.2037 - val_dense_78_loss: 0.3010 - val_dense_76_f1_m: 0.7930 - val_dense_76_f1_m_1: 0.7930 - val_dense_76_f1_m_2: 0.7930 - val_dense_77_f1_m: 0.9674 - val_dense_77_f1_m_1: 0.9674 - val_dense_77_f1_m_2: 0.9674 - val_dense_78_f1_m: 0.9364 - val_dense_78_f1_m_1: 0.9364 - val_dense_78_f1_m_2: 0.9364\n",
            "Epoch 48/50\n",
            "596/596 [==============================] - 0s 162us/step - loss: 0.7664 - dense_76_loss: 0.1460 - dense_77_loss: 0.3886 - dense_78_loss: 0.2318 - dense_76_f1_m: 0.7883 - dense_76_f1_m_1: 0.7883 - dense_76_f1_m_2: 0.7883 - dense_77_f1_m: 0.9555 - dense_77_f1_m_1: 0.9555 - dense_77_f1_m_2: 0.9555 - dense_78_f1_m: 0.8537 - dense_78_f1_m_1: 0.8537 - dense_78_f1_m_2: 0.8537 - val_loss: 0.7344 - val_dense_76_loss: 0.1655 - val_dense_77_loss: 0.3165 - val_dense_78_loss: 0.2524 - val_dense_76_f1_m: 0.5915 - val_dense_76_f1_m_1: 0.5915 - val_dense_76_f1_m_2: 0.5915 - val_dense_77_f1_m: 0.9696 - val_dense_77_f1_m_1: 0.9696 - val_dense_77_f1_m_2: 0.9696 - val_dense_78_f1_m: 0.8333 - val_dense_78_f1_m_1: 0.8333 - val_dense_78_f1_m_2: 0.8333\n",
            "Epoch 49/50\n",
            "596/596 [==============================] - 0s 159us/step - loss: 0.8488 - dense_76_loss: 0.2241 - dense_77_loss: 0.3277 - dense_78_loss: 0.2970 - dense_76_f1_m: 0.7466 - dense_76_f1_m_1: 0.7466 - dense_76_f1_m_2: 0.7466 - dense_77_f1_m: 0.9496 - dense_77_f1_m_1: 0.9496 - dense_77_f1_m_2: 0.9496 - dense_78_f1_m: 0.8696 - dense_78_f1_m_1: 0.8696 - dense_78_f1_m_2: 0.8696 - val_loss: 0.6101 - val_dense_76_loss: 0.1015 - val_dense_77_loss: 0.2745 - val_dense_78_loss: 0.2341 - val_dense_76_f1_m: 0.7863 - val_dense_76_f1_m_1: 0.7863 - val_dense_76_f1_m_2: 0.7863 - val_dense_77_f1_m: 0.9774 - val_dense_77_f1_m_1: 0.9774 - val_dense_77_f1_m_2: 0.9774 - val_dense_78_f1_m: 0.8590 - val_dense_78_f1_m_1: 0.8590 - val_dense_78_f1_m_2: 0.8590\n",
            "Epoch 50/50\n",
            "596/596 [==============================] - 0s 160us/step - loss: 0.6532 - dense_76_loss: 0.1287 - dense_77_loss: 0.2837 - dense_78_loss: 0.2407 - dense_76_f1_m: 0.8193 - dense_76_f1_m_1: 0.8193 - dense_76_f1_m_2: 0.8193 - dense_77_f1_m: 0.9577 - dense_77_f1_m_1: 0.9577 - dense_77_f1_m_2: 0.9577 - dense_78_f1_m: 0.8683 - dense_78_f1_m_1: 0.8683 - dense_78_f1_m_2: 0.8683 - val_loss: 0.5127 - val_dense_76_loss: 0.0742 - val_dense_77_loss: 0.1951 - val_dense_78_loss: 0.2434 - val_dense_76_f1_m: 0.9238 - val_dense_76_f1_m_1: 0.9238 - val_dense_76_f1_m_2: 0.9238 - val_dense_77_f1_m: 0.9776 - val_dense_77_f1_m_1: 0.9776 - val_dense_77_f1_m_2: 0.9776 - val_dense_78_f1_m: 0.9190 - val_dense_78_f1_m_1: 0.9190 - val_dense_78_f1_m_2: 0.9190\n",
            "(420, 1)\n",
            "(420, 1)\n",
            "(420, 5)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "casT5v9RXYa3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "db_wres = {}\n",
        "db_wres = {'yy1':yy1_w,'yy2':yy2_w,'yy3':yy3_w,'history':his_w}\n",
        "feed_data_to_pkl('weighted_res_hist',db_wres)\n",
        "feed_data_to_pkl('weighted_model',model_w)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m_4IyYtJLZc9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y3_test_f =[]\n",
        "y3_train_f = []\n",
        "for i in range(len(y3_test)):\n",
        "  y3_test_f.append(np.argmax(y3_test[i]))\n",
        "for i in range(len(y3_train)):\n",
        "  y3_train_f.append(np.argmax(y3_train[i])) \n",
        "y3_train_f = np.array(y3_train_f)\n",
        "y3_test_f = np.array(y3_test_f)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jhRj8gyGKVyr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "yy1_b,yy2_b,yy3_b,his_bal,model_bal = fit_predict_balanced_model(X_train, y1_train.ravel(),y2_train.ravel(),\n",
        "                                                     y3_train_f, X_test, y1_test.ravel(),y2_test.ravel(),y3_test_f)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "LpJczJi2Zy04",
        "colab_type": "code",
        "outputId": "be339f4d-affb-43c5-f3df-11e7647beb9f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "print(y3_train.shape)\n",
        "print(y3_test.shape)"
      ],
      "execution_count": 95,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(852, 5)\n",
            "(420, 5)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QvOU_KV6Oqm5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "0koQHAP6Zy1O",
        "colab_type": "code",
        "outputId": "6300fde9-3797-4162-93c3-c423f701a2f2",
        "colab": {}
      },
      "source": [
        "# evaluate the model\n",
        "f1score = model.evaluate(X_test, [y1_test,y2_test,y3_test])\n",
        "#classifier =KerasClassifier(build_fn=making_model(193), epochs=100, batch_size=10, verbose=0)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "410/410 [==============================] - 0s 100us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "6-dbhek0Zy1R",
        "colab_type": "code",
        "outputId": "418fdbdb-bd4b-4bde-c948-6c2ef4796fc5",
        "colab": {}
      },
      "source": [
        "f1score"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.4752868788998301,\n",
              " 0.08234787732362747,\n",
              " 0.10622140020132065,\n",
              " 0.2854807674884796,\n",
              " 0.7038729786872864,\n",
              " 0.9776144623756409,\n",
              " 0.9295616149902344]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "R-Bot_SNZy1W",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "yy1,yy2,yy3 = model.predict(X_test)\n",
        "def normalise_op(yy1,yy2,yy3):\n",
        "  shape = yy3.shape\n",
        "  for i in range(0, shape[0]):\n",
        "    for j in range(0, shape[1]):\n",
        "      if yy3[i, j] >= 0.5:\n",
        "        yy3[i, j] = 1\n",
        "      else:\n",
        "        yy3[i, j] = 0\n",
        "  for i in range(0, shape[0]):\n",
        "    if yy2[i, 0] >= 0.5:\n",
        "      yy2[i, 0] = 1\n",
        "    else:\n",
        "      yy2[i, 0] = 0\n",
        "  for i in range(0, shape[0]):\n",
        "    if yy1[i, 0] >= 0.5:\n",
        "      yy1[i, 0] = 1\n",
        "    else:\n",
        "      yy1[i, 0] = 0  \n",
        "  return yy1,yy2,yy3\n",
        "yy1,yy2,yy3 = normalise_op(yy1,yy2,yy3)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "iyZoEg66Zy1Z",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "res_pred=np.empty(shape=(len(yy3),1),dtype=int)\n",
        "res_ex = np.empty(shape=(len(yy3),1),dtype=int)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "0kszRdbmZy1c",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "for i in range(len(yy3)):\n",
        "    res_pred[i] = argmax(yy3[i])\n",
        "    res_ex[i] = argmax(y3_test[i])\n",
        "    #print(argmax(yy3[i]))\n",
        "    #print(y3_test[i])\n",
        "    #print(\"--------------------------------------------\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "G99HuFWBZy1e",
        "colab_type": "code",
        "outputId": "927e4c38-d590-4c49-d9aa-63a608affccd",
        "colab": {}
      },
      "source": [
        "print(confusion_matrix(y1_test,yy1))\n",
        "print(classification_report(y1_test,yy1))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[374   1]\n",
            " [ 10  25]]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.97      1.00      0.99       375\n",
            "           1       0.96      0.71      0.82        35\n",
            "\n",
            "    accuracy                           0.97       410\n",
            "   macro avg       0.97      0.86      0.90       410\n",
            "weighted avg       0.97      0.97      0.97       410\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "ZmdDCgQBZy1h",
        "colab_type": "code",
        "outputId": "106cb1a9-8b24-4ee0-fa76-8febef35e851",
        "colab": {}
      },
      "source": [
        "print(confusion_matrix(y2_test,yy2))\n",
        "print(classification_report(y2_test,yy2))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[ 14   8]\n",
            " [  9 379]]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.61      0.64      0.62        22\n",
            "           1       0.98      0.98      0.98       388\n",
            "\n",
            "    accuracy                           0.96       410\n",
            "   macro avg       0.79      0.81      0.80       410\n",
            "weighted avg       0.96      0.96      0.96       410\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "M-wNbOaTZy1j",
        "colab_type": "code",
        "outputId": "ce79b42e-42c8-4290-b253-8d8305018b99",
        "colab": {}
      },
      "source": [
        "confusion_matrix(res_ex,res_pred)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[371,   0,   0,   6,   0],\n",
              "       [ 13,   0,   0,   0,   0],\n",
              "       [ 17,   0,   0,   1,   0],\n",
              "       [  0,   0,   0,   0,   0],\n",
              "       [  2,   0,   0,   0,   0]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "lP6vd4aqZy1m",
        "colab_type": "code",
        "outputId": "d00b126d-bc83-4caa-ab01-94d0ee53c454",
        "colab": {}
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model_1\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_1 (InputLayer)            (None, 193)          0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential_1 (Sequential)       (None, 193)          0           input_1[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dense_1 (Dense)                 (None, 40)           7760        sequential_1[1][0]               \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_1 (BatchNor (None, 40)           160         dense_1[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dropout_1 (Dropout)             (None, 40)           0           batch_normalization_1[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "dense_2 (Dense)                 (None, 20)           820         dropout_1[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_2 (BatchNor (None, 20)           80          dense_2[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dropout_2 (Dropout)             (None, 20)           0           batch_normalization_2[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "dense_3 (Dense)                 (None, 8)            168         dropout_2[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_3 (BatchNor (None, 8)            32          dense_3[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dropout_3 (Dropout)             (None, 8)            0           batch_normalization_3[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "dense_4 (Dense)                 (None, 1)            9           dropout_3[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_5 (Dense)                 (None, 1)            9           dropout_3[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_6 (Dense)                 (None, 5)            45          dropout_3[0][0]                  \n",
            "==================================================================================================\n",
            "Total params: 9,083\n",
            "Trainable params: 8,947\n",
            "Non-trainable params: 136\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "4HsxCSq3Zy1o",
        "colab_type": "code",
        "outputId": "8a22a87e-0740-4b00-d2c6-f0ef7ad51e4b",
        "colab": {}
      },
      "source": [
        "print(f1_score(res_ex,res_pred,average='weighted'))\n",
        "print(precision_score(res_ex,res_pred,average='weighted'))\n",
        "print(recall_score(res_ex,res_pred,average='weighted'))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.8747154471544717\n",
            "0.8464988198269079\n",
            "0.9048780487804878\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/opt/conda/lib/python3.6/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
            "  'precision', 'predicted', average, warn_for)\n",
            "/opt/conda/lib/python3.6/site-packages/sklearn/metrics/classification.py:1439: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no true samples.\n",
            "  'recall', 'true', average, warn_for)\n",
            "/opt/conda/lib/python3.6/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
            "  'precision', 'predicted', average, warn_for)\n",
            "/opt/conda/lib/python3.6/site-packages/sklearn/metrics/classification.py:1439: UndefinedMetricWarning: Recall is ill-defined and being set to 0.0 in labels with no true samples.\n",
            "  'recall', 'true', average, warn_for)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "QgWDA_e0Zy1s",
        "colab_type": "code",
        "outputId": "1c5d292d-5a22-4b4d-8b5e-8a7a8718ebb9",
        "colab": {}
      },
      "source": [
        "print(f1_score(res_ex,res_pred,average='micro'))\n",
        "print(precision_score(res_ex,res_pred,average='micro'))\n",
        "print(recall_score(res_ex,res_pred,average='micro'))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9048780487804878\n",
            "0.9048780487804878\n",
            "0.9048780487804878\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "kBaOW2W3Zy1v",
        "colab_type": "code",
        "outputId": "346d37f7-99a1-4563-a659-3ab237adcda4",
        "colab": {}
      },
      "source": [
        "accuracy_score(res_ex,res_pred)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9048780487804878"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "WxdXsxJ5Zy1y",
        "colab_type": "code",
        "outputId": "22a2c6d3-9148-4441-fcb0-266122b541b9",
        "colab": {}
      },
      "source": [
        "print(f1_score(res_ex,res_pred,average='macro'))\n",
        "print(precision_score(res_ex,res_pred,average='macro'))\n",
        "print(recall_score(res_ex,res_pred,average='macro'))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.19025641025641027\n",
            "0.18411910669975187\n",
            "0.19681697612732096\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "yCu81Ic2Zy10",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from imblearn.keras import BalancedBatchGenerator\n",
        "from sklearn.metrics import roc_auc_score\n",
        "\n",
        "def fit_predict_balanced_model(X_train, y_train, X_test, y_test):\n",
        "    model = classifier\n",
        "    training_generator = BalancedBatchGenerator(X_train, y_train)\n",
        "    model.fit_generator(generator=training_generator,epochs=100,verbose=1)\n",
        "    y_pred = model.predict_proba(X_test)\n",
        "    print(classification_report(y_test, y_pred))\n",
        "    return roc_auc_score(y_test, y_pred)\n",
        "res = fit_predict_balanced_model(X_train, y_train, X_test, y_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "olPs6zgJZy12",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "classifier"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "trusted": true,
        "id": "3OAt4Af-Zy15",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "7qAVm6BttvHG",
        "trusted": true,
        "colab": {}
      },
      "source": [
        "# Fitting our model \n",
        "import eli5\n",
        "from eli5.sklearn import PermutationImportance\n",
        "\n",
        "history=model.fit(X_train, y_train,validation_data=(X_test, y_test),nb_epoch = 100)\n",
        "#history=classifier.fit(X,y,validation_data=(X_test, y_test), batch_size = 10, nb_epoch = 100)\n",
        "#acc = []\n",
        "# Predicting the Test set results"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "4iXz1qF3tvJY",
        "trusted": true,
        "colab": {}
      },
      "source": [
        "# F1 score\n",
        "f1_score(y_test,y_pred)"
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}