predictor = Predictor(output=['z'])
def iter_function(epoch, error, test_error, test_error_gradient, test_accuracy):
print(
'epoch: {iter}, error: {error}, test_error: {test_error}, test_error_gradient: {test_error_gradient}, test_accuracy: {test_accuracy}'.format(
iter=epoch, error=error, test_error=test_error, test_error_gradient=test_error_gradient,
accuracy=predictor.train_accuracy, test_accuracy=test_accuracy))
predictor.learn(from_data=data_frame, callback_on_iter=iter_function)
print('accuracy')
print(predictor.train_accuracy)
print('accuracy over all dataset')
print(predictor.calculate_accuracy(from_data=data_frame))
when = {'x': [1], 'y': [0]}
print('- multiply when. {when}'.format(when=when))
print(predictor.predict(when=when))
# saving the predictor
predictor.save('ok.pkl')
# loading the predictor
predictor2 = Predictor(load_from_path='ok.pkl')
when = {'x': [0, 0, 1, -1, 1], 'y': [0, 1, -1, -1, 1]}
print('- multiply when. {when}'.format(when=when))
print(predictor2.predict(when_data=pandas.DataFrame(when)))
when = {'x': [0, 3, 1, -5, 1], 'y': [0, 1, -5, -4, 7]}
print('- multiply when. {when}'.format(when=when))
def iter_function(epoch, training_error, test_error, test_error_gradient,
test_accuracy):
print(
f'Epoch: {epoch}, Train Error: {training_error}, Test Error: {test_error}, Test Error Gradient: {test_error_gradient}, Test Accuracy: {test_accuracy}'
)
predictor.learn(from_data=df_train,
callback_on_iter=iter_function,
eval_every_x_epochs=200)
predictor.save('ok.pkl')
predictor = Predictor(load_from_path='ok.pkl')
print('Train accuracy: ', predictor.train_accuracy)
print('Test accuracy: ', predictor.calculate_accuracy(from_data=df_test))
predictions = predictor.predict(when_data=df_test)
print('Confidence mean for both x and y present: ',
np.mean(predictions['z']['selfaware_confidences']))
print(list(df_test['z'])[30:60])
print(predictions['z']['predictions'][30:60])
predictions = predictor.predict(when_data=df_test.drop(columns=['x']))
print('Confidence mean for x missing: ',
np.mean(predictions['z']['selfaware_confidences']))
predictions = predictor.predict(when_data=df_test.drop(columns=['y']))
print('Confidence mean for y missing: ',
np.mean(predictions['z']['selfaware_confidences']))
