Human-Activity-Detection

In this module we trained a CNN to recognize different human activities performed in real life from raw accelerometer signals collected from a smartphone. The model consists of one convolution layer followed by max pooling and another convolution layer. After that, the model will have fully connected layer which is connected to Softmax layer.

After running it for 5 epochs we achieve:

Training Loss: 0.7445856
Training Accuracy: 0.871372

DataSet from WISDM: WIreless Sensor Data Mining

The Actitracker data set released by Wireless Sensor Data Mining (WISDM) lab.

http://www.cis.fordham.edu/wisdm/dataset.php

DataSet Description

This dataset contains six daily activities collected in a controlled laboratory environment.
The activities include jogging, walking, ascending stairs, descending stairs, sitting and standing.
The data is collected from 36 users using a smartphone in their pocket with the 20Hz sampling rate (20 values per second).

Statistics

Raw Time Series Data

Number of examples: 1,098,207
Number of attributes: 6
Missing attribute values: None

Class Distribution

Walking: 424,400 (38.6%)
Jogging: 342,177 (31.2%)
Upstairs: 122,869 (11.2%)
Downstairs: 100,427 (9.1%)
Sitting: 59,939 (5.5%)
Standing: 48,395 (4.4%)