X_test = scaler.fit_transform(X_test)
#
clf = RandomForestClassifier(n_estimators=n,
max_features=m,
n_jobs=6,
criterion=c)
clf.fit(X_train, y_train)
#
# Make predictions for validation data and evaluate
pred_y = clf.predict(X_test)
#
# Make predictions for testing data and evaluate
pred_y2 = clf.predict(df_pruned_shifted_X2)
#
# Testing Classifier Accuracy on Verification Dataset
sf2 = Scoring_Functions(y_pred=pred_y2,
y_true=df_pruned_shifted_Y2)
#
# Testing Classifier Accuracy on Verification Dataset
sf = Scoring_Functions(y_pred=pred_y, y_true=y_test)
count += 1
if sf2.accuracy() > 60 and sf2.f_measure() > 60:
print("(" + str(count) + ") Criterion: " + str(c) +
"\nn_estimators: " + str(n) + "\nmax_features: " +
str(m) + "\ntest_size: " + str(ts) +
"\n------------")
print("Verification Sample:")
print(sf.scoring_results())
print("------------")
print("Test Sample:")
print(sf2.scoring_results())
print(
X_test = scaler.fit_transform(X_test)
#
# using a grid search to find optimum hyper parameter
from sklearn import svm
from sklearn.model_selection import GridSearchCV
parameters = {
'C': (1, 2, 3, 4, 5, 6, 7),
'gamma': [40, 35, 30, 27, 25, 23, 20]
}
clf = svm.SVC()
clf = GridSearchCV(clf, parameters)
clf.fit(X_train, y_train)
print(clf.best_params_)
kernel = 'rbf'
C = clf.best_params_['C']
gamma = clf.best_params_['gamma']
degree = 3
clf = svm.SVC(kernel=kernel, C=C, gamma=gamma, degree=degree)
clf.fit(X_train, y_train)
print(clf)
#
# make predictions for test data and evaluate
pred_y = clf.predict(X_test)
#
# Testing Classifier Accuracy
from src.statistics.scoring_functions import Scoring_Functions
sf = Scoring_Functions(y_pred=pred_y, y_true=y_test)
print("SVM Accuracy: ")
print(sf.scoring_results())
print('-------------------------')
X_train, X_test, y_train, y_test = train_test_split(
df_pruned_shifted_X,
df_pruned_shifted_Y,
test_size=ts,
random_state=0)
X_train = scaler.fit_transform(X_train)
X_test = scaler.fit_transform(X_test)
#
clf = svm.SVC(kernel=kernel, C=c, gamma=g, degree=degree)
clf.fit(X_train, y_train)
#
# Make predictions for validation data and evaluate
pred_y = clf.predict(X_test)
#
# Testing Classifier Accuracy on Verification Dataset
sf = Scoring_Functions(y_pred=pred_y, y_true=y_test)
print("Gamma: " + str(g) + "\nC: " + str(c) + "\ntest_size: " +
str(ts) + "\n------------")
print("Verification Sample:")
print(sf.scoring_results())
#
# Make predictions for testing data and evaluate
pred_y = clf.predict(df_pruned_shifted_X2)
#
# Testing Classifier Accuracy on Verification Dataset
sf = Scoring_Functions(y_pred=pred_y, y_true=df_pruned_shifted_Y2)
print("------------")
print("Test Sample:")
print(sf.scoring_results())
print(
"----------------------------------------------------------------------"