X_train = features[:Ntrain]
X_test = features[Ntrain:]
y_train = label[:Ntrain]
y_test = label[Ntrain:]
############################################################################
# simulate active learning process
accuracy = []
Nrecs = 200; # number of new labels in each iteration
for k in range(6):
# 1. train model with training data, and recommend samples for labeling
out = AL.process(X_train, y_train, X_test, Nrecs, 'SVM')
IDtoLabel = out['IDtoLabel']
relevanceScore = out['relevanceScore']
# 2. assume get the expert feedback
y_newlabel = y_test[IDtoLabel]
# For Simulation purpose
# performance evaluation: calculate prediction accuracy
classifier = out['classifier']
y_test_predicted = classifier.predict(X_test)
out = evaluate(y_test, y_test_predicted)
accuracy.append(out['accuracy'])
# updated y_test
X_train = features[:Ntrain]
X_test = features[Ntrain:]
y_train = label[:Ntrain]
y_test = label[Ntrain:]
############################################################################
# simulate active learning process
accuracy = []
Nrecs = 50; # number of new labels in each iteration
for k in range(6):
# 1. train model with training data, and recommend samples for labeling
out = AL.process(X_train, y_train, X_test, Nrecs, 'SVM', 'entropy')
IDtoLabel = out['IDtoLabel']
relevanceScore = out['relevanceScore']
# 2. assume get the expert feedback
y_newlabel = y_test[IDtoLabel]
# For Simulation purpose
# performance evaluation: calculate prediction accuracy
classifier = out['classifier']
y_test_predicted = classifier.predict(X_test)
out = evaluate(y_test, y_test_predicted)
accuracy.append(out['accuracy'])
# updated y_test
