def svm_loan(train_dataset, test_dataset, eval_dataset, selected_features):
X_test = test_dataset.drop(columns=["status"])
y_test = test_dataset.iloc[:, -1]
X_train = train_dataset.drop(columns=["status"])
y_train = train_dataset.iloc[:, -1]
X_train = X_train[selected_features]
X_test = X_test[selected_features]
X_test = X_test.values
y_test = y_test.values
X_train = X_train.values
y_train = y_train.values
X_train, y_train = smote_sampling(X_train, y_train)
scaler = StandardScaler()
scaler.fit(X_train)
X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)
#svclassifier = SVC(gamma='auto')
#svclassifier.fit(X_train, y_train)
svclassifier = parameters_tuner(X_train, y_train)
y_pred = svclassifier.predict(X_test)
predictions = [round(value) for value in y_pred]
accuracy = accuracy_score(y_test, predictions)
print(str(confusion_matrix(y_test, y_pred)))
print(str(classification_report(y_test, y_pred, zero_division=0)))
print(f"AUC: {roc_auc_score(y_test, y_pred)}")
print("Accuracy: %.2f%%" % (accuracy * 100.0))
plot_auc(svclassifier, X_test, y_test, "SVM")
X_eval = eval_dataset.drop(columns=["status"])
X_eval = X_eval[selected_features].values
X_eval = scaler.transform(X_eval)
id_array = map(lambda x: int(x), eval_dataset.index.values)
y_pred = map(lambda x: int(x), svclassifier.predict(X_eval))
result = pd.DataFrame({
'Id': id_array,
'Predicted': y_pred
})
return result
def xg_boost(train_dataset, test_dataset, eval_dataset, selected_features):
#train_dataset = down_sampling(train_dataset)
X_test = test_dataset.drop(columns=["status"])
y_test = test_dataset.iloc[:, -1]
X_train = train_dataset.drop(columns=["status"])
y_train = train_dataset.iloc[:, -1]
X_train = X_train[selected_features]
X_test = X_test[selected_features]
X_test = X_test.values
y_test = y_test.values
X_train = X_train.values
y_train = y_train.values
X_train, y_train = smote_sampling(X_train, y_train)
scaler = StandardScaler()
scaler.fit(X_train)
X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)
model = XGBClassifier()
model.fit(X_train, y_train)
y_pred = model.predict(X_test)
predictions = [round(value) for value in y_pred]
accuracy = accuracy_score(y_test, predictions)
current_AUC = roc_auc_score(y_test, y_pred)
print(str(confusion_matrix(y_test, y_pred)))
print(str(classification_report(y_test, y_pred, zero_division=0)))
print(f"Current AUC: {current_AUC}")
print("Accuracy: %.2f%%" % (accuracy * 100.0))
plot_auc(model, X_test, y_test, "XGBoost")
X_eval = eval_dataset.drop(columns=["status"])
X_eval = X_eval[selected_features].values
X_eval = scaler.transform(X_eval)
id_array = map(lambda x: int(x), eval_dataset.index.values)
y_pred = map(lambda x: int(x), model.predict(X_eval))
result = pd.DataFrame({'Id': id_array, 'Predicted': y_pred})
return result
def crforest_loan(train_dataset, test_dataset, eval_dataset,
selected_features):
X_test = test_dataset.drop(columns=["status"])
y_test = test_dataset.iloc[:, -1]
X_train = train_dataset.drop(columns=["status"])
y_train = train_dataset.iloc[:, -1]
X_train = X_train[selected_features]
X_test = X_test[selected_features]
X_test = X_test.values
y_test = y_test.values
X_train = X_train.values
y_train = y_train.values
X_train, y_train = smote_sampling(X_train, y_train)
scaler = StandardScaler()
scaler.fit(X_train)
X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)
#clf = RandomForestClassifier(max_depth=2, random_state=0)
#clf.fit(X_train, y_train)
clf = parameters_tuner(X_train, y_train)
y_pred = clf.predict(X_test)
predictions = [round(value) for value in y_pred]
accuracy = accuracy_score(y_test, predictions)
current_AUC = roc_auc_score(y_test, y_pred)
print(str(confusion_matrix(y_test, y_pred)))
print(str(classification_report(y_test, y_pred, zero_division=0)))
print(f"Current AUC: {current_AUC}")
print("Accuracy: %.2f%%" % (accuracy * 100.0))
plot_auc(clf, X_test, y_test, "random_forest")
X_eval = eval_dataset.drop(columns=["status"])
X_eval = X_eval[selected_features].values
X_eval = scaler.transform(X_eval)
id_array = map(lambda x: int(x), eval_dataset.index.values)
y_pred = map(lambda x: int(x), clf.predict(X_eval))
result = pd.DataFrame({'Id': id_array, 'Predicted': y_pred})
return result
def gbm(train_dataset, test_dataset, eval_dataset, selected_features):
X_test = test_dataset.drop(columns=["status"])
y_test = test_dataset.iloc[:, -1]
X_train = train_dataset.drop(columns=["status"])
y_train = train_dataset.iloc[:, -1]
X_train = X_train[selected_features]
X_test = X_test[selected_features]
X_test = X_test.values
y_test = y_test.values
X_train = X_train.values
y_train = y_train.values
scaler = StandardScaler()
# X_train, y_train = .fit_resample(X_train, y_train)
scaler.fit(X_train)
X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)
gbc = GradientBoostingClassifier(max_depth=15, random_state=0)
gbc.fit(X_train, y_train)
y_pred = gbc.predict(X_test)
predictions = [round(value) for value in y_pred]
accuracy = accuracy_score(y_test, predictions)
print(str(confusion_matrix(y_test, y_pred)))
print(str(classification_report(y_test, y_pred, zero_division=0)))
print(f"AUC: {roc_auc_score(y_test, y_pred)}")
print("Accuracy: %.2f%%" % (accuracy * 100.0))
plot_auc(gbc, X_test, y_test, "gbm")
X_eval = eval_dataset.drop(columns=["status"])
X_eval = X_eval[selected_features].values
X_eval = scaler.transform(X_eval)
id_array = map(lambda x: int(x), eval_dataset.index.values)
y_pred = map(lambda x: int(x), gbc.predict(X_eval))
result = pd.DataFrame({'Id': id_array, 'Predicted': y_pred})
return result
def knn_loan(train_dataset, test_dataset, eval_dataset, selected_features):
X_test = test_dataset.drop(columns=["status"])
y_test = test_dataset.iloc[:, -1]
X_train = train_dataset.drop(columns=["status"])
y_train = train_dataset.iloc[:, -1]
X_train = X_train[selected_features]
X_test = X_test[selected_features]
X_test = X_test.values
y_test = y_test.values
X_train = X_train.values
y_train = y_train.values
X_train, y_train = smote_sampling(X_train, y_train)
scaler = StandardScaler()
scaler.fit(X_train)
X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)
auc = []
# Calculating error for K values between 1 and 40
for i in range(1, 40):
knn = KNeighborsClassifier(n_neighbors=i)
knn.fit(X_train, y_train)
pred_i = knn.predict(X_test)
auc.append(roc_auc_score(y_test, pred_i))
best_k = auc.index(max(auc)) + 1
fpr, tpr, _thresholds = metrics.roc_curve(y_test, pred_i)
print(f"AUC: {metrics.auc(fpr, tpr)}")
knn = KNeighborsClassifier(n_neighbors=best_k)
knn.fit(X_train, y_train)
y_pred = knn.predict(X_test)
print(str(confusion_matrix(y_test, y_pred)))
print(str(classification_report(y_test, y_pred, zero_division=0)))
print(f"AUC: {roc_auc_score(y_test, y_pred)}")
print(f"Best K: {best_k}")
plot_auc(knn, X_test, y_test, "knn")
X_eval = eval_dataset.drop(columns=["status"])
X_eval = X_eval[selected_features].values
X_eval = scaler.transform(X_eval)
id_array = map(lambda x: int(x), eval_dataset.index.values)
y_pred = map(lambda x: int(x), knn.predict(X_eval))
result = pd.DataFrame({
'Id': id_array,
'Predicted': y_pred
})
return result