-
Notifications
You must be signed in to change notification settings - Fork 0
/
demo_simulated.py
46 lines (35 loc) · 1.73 KB
/
demo_simulated.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
from time import time
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from models.svm import CustomSVM
from utils import make_dataset, plot_misclassification_error
# Generate a binary dataset with 200 datapoints and 20 features
random_state = 0 # set random seed to allow reproducing results
X, y = make_dataset(random_state=random_state)
# Split the dataset into train and test in the ratio 80-20
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2,
random_state=random_state)
# Standardize the X values for both the train and the test data
scaler = StandardScaler().fit(X_train)
X_train_scaled = scaler.transform(X_train)
X_test_scaled = scaler.transform(X_test)
# Initialize the CustomSVM classifier with default parameters
clf = CustomSVM()
# Fit the model on the training data and time the process
print('Training the model...')
start = time()
clf.fit(X_train_scaled, y_train)
end = time()
print('Training completed.')
# Calculate the accuracy of the model on the training and the testing data
train_score = clf.score(X_train_scaled, y_train)
test_score = clf.score(X_test_scaled, y_test)
# Print the performance metrics of the model
print('\nPerformance metrics:')
print('Time taken to fit the model: {0:.4f}s'.format(end - start))
print('Accuracy on the training data: {0:.4f}'.format(train_score))
print('Accuracy on the testing data: {0:.4f}'.format(test_score))
# Plot the misclassification error and save the image to a file
output_file = './misclassification_error_simulated_custom_svm.png'
print('Saving misclassification plot to ' + output_file)
plot_misclassification_error(clf, X_train_scaled, y_train, output_file)