def torax():
print("Started training classifiers on torax data set.")
x_train, x_test, y_train, y_test = data_processing.torax()
result = {}
# LR
lr_best_model, lr_params = classification_cv.logistic_regression(
x_train, y_train, C_min=0.01, C_max=10, fold=3, iterations=20)
result[constant.LR] = evaluate_classifier(x_train, x_test, y_train, y_test,
lr_best_model, lr_params,
"torax", "Logistic Regression")
# DECISION TREE
dt_best_model, dt_params = classification_cv.decision_trees(x_train,
y_train,
max_depth=10,
fold=3,
iterations=20)
result[constant.DECISION_TREE] = evaluate_classifier(
x_train, x_test, y_train, y_test, dt_best_model, dt_params, "torax",
"DECISION TREE")
# RANDOM FOREST
rf_best_model, rf_params = classification_cv.random_forest(
x_train, y_train, max_estimator=20, fold=4, iterations=20)
result[constant.RANDOM_FOREST] = evaluate_classifier(
x_train, x_test, y_train, y_test, rf_best_model, rf_params, "torax",
"RANDOM FOREST")
# SVC
# svc_best_model, svc_params = classification_cv.SVC(x_train, y_train,
# ['linear','poly', 'rbf', 'sigmoid'],
# 0.01, 100, 1, 1000, fold=2,
# iterations=20)
# evaluate_classifier(x_train, x_test, y_train, y_test, svc_best_model, svc_params, "torax", "SVC")
# KNN
knn_best_model, knn_params = classification_cv.KNC(x_train,
y_train,
neighbors=10,
fold=4,
iterations=20)
result[constant.KNC] = evaluate_classifier(x_train, x_test, y_train,
y_test, knn_best_model,
knn_params, "torax", "KNN")
# GaussianNB
nb_best_model, nb_params = classification_cv.GaussianNB(x_train,
y_train,
fold=4,
iterations=20)
result[constant.GNB] = evaluate_classifier(x_train, x_test, y_train,
y_test, nb_best_model,
nb_params, "torax",
"Gaussian NB")
# Ada boost classifier
AB_best_model, AB_params = classification_cv.ada_boost_classifier(
x_train, y_train, no_estimators=100, fold=4, iterations=20)
result[constant.ADABOOST] = evaluate_classifier(x_train, x_test, y_train,
y_test, AB_best_model,
AB_params, "torax",
"ADA BOOST")
# MLP
mlp_best_model, mlp_params = classification_cv.MLPClassifier(
x_train,
y_train,
hidden_layer_sizes=[(10, ), (
100,
10,
)],
alphas=[0.01, 0.05, 0.5, 1],
max_iter=[10, 20, 50, 100],
fold=4,
iterations=20)
result[constant.NN] = evaluate_classifier(x_train, x_test, y_train, y_test,
mlp_best_model, mlp_params,
"torax", "MLP")
export_result(result, "result/classification/torax.json")
def default_credit_card():
print(
"Started training classifiers on default credit card clients data set."
)
x_train, x_test, y_train, y_test = data_processing.credit_card_client()
result = {}
# LR
lr_best_model, lr_params = classification_cv.logistic_regression(
x_train, y_train, C_min=0.01, C_max=10, fold=3, iterations=20)
result[constant.LR] = evaluate_classifier(x_train, x_test, y_train, y_test,
lr_best_model, lr_params,
"Default credit card",
"Logistic Regression")
# DECISION TREE
dt_best_model, dt_params = classification_cv.decision_trees(x_train,
y_train,
max_depth=10,
fold=3,
iterations=20)
result[constant.DECISION_TREE] = evaluate_classifier(
x_train, x_test, y_train, y_test, dt_best_model, dt_params,
"Default credit card", "DECISION TREE")
# RANDOM FOREST
rf_best_model, rf_params = classification_cv.random_forest(
x_train, y_train, max_estimator=20, fold=4, iterations=20)
result[constant.RANDOM_FOREST] = evaluate_classifier(
x_train, x_test, y_train, y_test, rf_best_model, rf_params,
"Default credit card", "RANDOM FOREST")
# AdaBoost
rf_best_model, rf_params = classification_cv.ada_boost_classifier(
x_train, y_train, no_estimators=50, fold=4, iterations=20)
result[constant.ADABOOST] = evaluate_classifier(x_train, x_test, y_train,
y_test, rf_best_model,
rf_params,
"Default credit card",
"AdaBoost")
# SVC
svc_best_model, svc_params = classification_cv.SVC(x_train,
y_train, ['linear'],
C_min=0.01,
C_max=1,
gamma_min=1,
gamma_max=100,
fold=3,
iterations=10)
result[constant.SVC] = evaluate_classifier(x_train, x_test, y_train,
y_test, svc_best_model,
svc_params,
"Default credit card", "SVC")
# KNN
knn_best_model, knn_params = classification_cv.KNC(x_train,
y_train,
neighbors=10,
fold=3,
iterations=20)
result[constant.KNC] = evaluate_classifier(x_train, x_test, y_train,
y_test, knn_best_model,
knn_params,
"Default credit card", "KNN")
# GaussianNB
nb_best_model, nb_params = classification_cv.GaussianNB(x_train,
y_train,
fold=4,
iterations=20)
result[constant.GNB] = evaluate_classifier(x_train, x_test, y_train,
y_test, nb_best_model,
nb_params,
"Default credit card",
"Gaussian NB")
# MLP
mlp_best_model, mlp_params = classification_cv.MLPClassifier(
x_train,
y_train,
hidden_layer_sizes=[(), (10, )],
alphas=[0.01, 0.05, 0.5, 1],
max_iter=[100, 1000, 10000, 100000],
fold=4,
iterations=20)
result[constant.NN] = evaluate_classifier(x_train, x_test, y_train, y_test,
mlp_best_model, mlp_params,
"Default credit card", "MLP")
export_result(result, "result/classification/default_credit_card.json")