def train_VotingClassifier():
data = pd.read_csv("processed_data.csv", sep=',')
data = shuffle(data)
data_X = data.drop('Empathy', axis=1)
data_Y = data['Empathy']
"""TRAINING"""
clf1 = joblib.load('SGD_model.pkl')
clf2 = joblib.load('SVC_model.pkl')
clf3 = joblib.load('RF_model.pkl')
# clf4 = joblib.load('Bernoulli_nb_model.pkl')
# clf5 = joblib.load('Multinomial_nb_model.pkl')
clf6 = joblib.load('LR_model.pkl')
clf7 = joblib.load('ET_model.pkl')
eclf1 = VotingClassifier(estimators=[('SGD', clf1),('SVC', clf2), ('RF', clf3), ('ET', clf7)], voting='hard')
eclf1.fit(data_X, data_Y)
joblib.dump(eclf1, 'Voting_model.pkl')
"""PERFORMANCE EVALUATION"""
accuracy, clf_report = get_cv_metrics(eclf1, data_X, data_Y, k_split=10)
print("Accuracy: ", accuracy) #Accuracy: 73.06
print(clf_report)
return eclf1
def train_LR():
data = pd.read_csv("processed_data.csv", sep=',')
data = shuffle(data)
data_X = data.drop('Empathy', axis=1)
data_Y = data['Empathy']
'''TRAINING'''
clf = LogisticRegression(C=2.6,
dual=False,
penalty='l2',
tol=1e-4,
solver='saga',
n_jobs=-1,
multi_class='ovr',
fit_intercept=True,
intercept_scaling=1.4).fit(
data_X, data_Y.values.ravel())
## clf = RFECV(clf, step=1, cv=5, n_jobs=-1).fit(data_X,data_Y)
joblib.dump(clf, 'LR_model.pkl')
"""PERFORMANCE EVALUATION"""
accuracy, clf_report = get_cv_metrics(clf,
data_X,
data_Y.values.ravel(),
k_split=10)
print("Accuracy: ", accuracy)
print(clf_report)
return clf
def train_SVM():
data = pd.read_csv("processed_data.csv", sep=',')
data = shuffle(data)
data_X = data.drop('Empathy', axis=1)
data_Y = data['Empathy']
'''TRAINING'''
clf = SVC(C=3.1, kernel='rbf', gamma=0.1,
tol=1e-1).fit(data_X, data_Y.values.ravel())
joblib.dump(clf, 'SVC_model.pkl')
"""PERFORMANCE EVALUATION"""
accuracy, clf_report = get_cv_metrics(clf,
data_X,
data_Y.values.ravel(),
k_split=10)
print("Accuracy: ", accuracy)
print(clf_report)
return clf
def train_RF():
data = pd.read_csv("processed_data.csv", sep=',')
data = shuffle(data)
data_X = data.drop('Empathy', axis=1)
data_Y = data['Empathy']
'''TRAINING'''
clf = RandomForestClassifier(n_estimators = 700, max_depth=300, random_state=16, n_jobs = -1)
## clf = RFECV(clf, step=1, cv=5, n_jobs=-1).fit(data_X,data_Y.values.ravel())
joblib.dump(clf, 'RF_model.pkl')
"""PERFORMANCE EVALUATION"""
accuracy, clf_report = get_cv_metrics(clf, data_X, data_Y.values.ravel(), k_split=10)
print("Accuracy: ", accuracy)
print(clf_report)
return clf
def train_BernoulliNB():
data = pd.read_csv("processed_data.csv", sep=',')
data = shuffle(data)
data_X = data.drop('Empathy', axis=1)
data_Y = data['Empathy']
'''TRAINING'''
clf = BernoulliNB(alpha=1.4, fit_prior=True, class_prior=None)
clf = RFECV(clf, step=1, cv=5, n_jobs=-1).fit(data_X, data_Y)
joblib.dump(clf, 'Bernoulli_nb_model.pkl')
"""PERFORMANCE EVALUATION"""
accuracy, clf_report = get_cv_metrics(clf,
data_X,
data_Y.values.ravel(),
k_split=10)
print("Accuracy: ", accuracy)
print(clf_report)
return clf