def __init__(self, file_path):
data_handler = DataHandler(file_path)
data_handler.remove_nans()
data_handler.categorical_to_num()
self.df = data_handler.data
def train(self, data):
X = data[self.cols].values
Y = data["Survived"].values
if self.classifier == 'logistic':
log_reg = LogisticRegression()
log_reg = log_reg.fit(X,Y)
score_log = cross_val_score(log_reg, X, Y, cv=5).mean()
print(score_log)
elif self.classifier == 'perceptron':
perceptron = Perceptron(
class_weight='balanced'
)
perceptron = perceptron.fit(X,Y)
score_pctr = cross_val_score(perceptron, X, Y, cv=5).mean()
print(score_pctr)
else:
raise ValueError("Unknown classifier.")
if __name__ == '__main__':
data_handler = DataHandler('./data/train.csv')
data_handler.remove_nans()
data_handler.categorical_to_num()
trainer1 = Trainer(["Age", "Sex"], 'logistic')
trainer1.train(data_handler.data)
trainer2 = Trainer(["Age", "Sex"], 'perceptron')
trainer2.train(data_handler.data)
