class CrossValidation:
"""
This class does k cross validation
"""
def __init__(self, model, hyperparameters, kfold):
self.metrics = Metrics()
cross_validation = StratifiedKFold(n_splits=kfold, shuffle=True)
self.clf = GridSearchCV(model, hyperparameters, cv=cross_validation, n_jobs=-1, verbose=1)
def fit_and_predict(self, x_train, y_train, x_test, y_test, metrics):
prediction = self.clf.fit(x_train, y_train).best_estimator_.predict(x_test)
if metrics == "accuracy":
self.metrics.accuracy(self.clf, y=y_test, pred=prediction)
elif metrics == "confusion_matrix":
self.metrics.confusion_matrix(self.clf, y=y_test, pred=prediction)
elif metrics == "roc":
prob = self.clf.fit(x_train, y_train).best_estimator_.predict_proba(x_test)
self.metrics.plot_roc(self.clf, y=y_test, prob=prob[:, 1])
def get_score(self, x_test, y_test):
return round(self.clf.score(x_test, y_test) * 100, 2)
class AbstractClassifier:
"""
Parent class of all project classifiers.
Attributes:
model : An object that defines the classifier model to implement.
metrics : An object that defines the different metrics that can be used to evaluate a model.
X_train : The features of the training data
Y_train : The targets of training data (the ground truth label)
X_test : The features of the testing data
Y_test : The targets of training data (the ground truth label)
"""
def __init__(self, model, mode='0'):
self.model = model
self.metrics = Metrics()
if mode == '0':
self.X_train, self.Y_train, self.X_test, self.Y_test = DataPreprocessing(
).naive_preprocessing_data()
elif mode == '1':
self.X_train, self.Y_train, self.X_test, self.Y_test = DataPreprocessing(
).advanced_preprocessing_data()
def train(self):
self.model.fit(self.X_train, self.Y_train)
def predict(self, x):
return self.model.predict(x)
def evaluate(self, label="Training", metrics="accuracy"):
if label == 'Training':
x, y = self.X_train, self.Y_train
else:
x, y = self.X_test, self.Y_test
if metrics == "accuracy":
self.metrics.accuracy(self.model, y, x, label)
elif metrics == "confusion_matrix":
self.metrics.confusion_matrix(self.model, y, x, label)
elif metrics == "roc":
self.metrics.plot_roc(self.model, y, x, label)
def tunning_model(self, hyperparameters, kfold, metrics):
cross_validate_model = CrossValidation(self.model, hyperparameters,
kfold)
cross_validate_model.fit_and_predict(self.X_train, self.Y_train,
self.X_test, self.Y_test, metrics)
return cross_validate_model.get_score(self.X_test, self.Y_test)