def TrainNetClassifier(Trained_labels, clusterAlg, df1Norm, fitmode, nclusters, prop):
x_Train, x_Test, y_Train, y_Test = train_test_split(df1Norm.values, Trained_labels, test_size=prop,
shuffle=False)
Features = len(x_Train[0])
clf = MLPClassifier(hidden_layer_sizes=(Features, Features, Features), max_iter=500, alpha=0.001,
solver='adam', random_state=21, tol=0.00000001)
predicted_cmeans = []
classesT = [i for i in range(nclusters)]
if fitmode.lower() == 'full' or fitmode == "":
try:
# experimental
if 'cmeans' in clusterAlg or 'c-means' in clusterAlg or 'fuzzy' in clusterAlg:
clf = clf.fit(x_Train, y_Train)
predicted_cmeans = clf.predict(x_Test)
else:
clf.fit(x_Train, y_Train)
except:
clf.fit(x_Train, y_Train)
else:
# experimental
for xa, ya in zip(x_Train, y_Train):
varw = ya.reshape(1, )
aarw = xa.reshape(1, -1)
clf._partial_fit(aarw, varw, classes=classesT)
return clf, predicted_cmeans, x_Test, x_Train, y_Test, y_Train
