def test_end_to_end(self):
model = RandomForestClassifier()
train_data = pd.read_csv("data/train.csv")
train_data = train_data[:300]
labels = train_data['label'].values
x = np.array(train_data.drop('label', axis=1))
y = labels
model.fit(data_preprocess(x), y)
y_predict = model.predict(data_preprocess(x))
self.assertGreater(f1_score(y, y_predict), 0)
split_size=split_size,
max_features=max_features)
RF.fit(X_train, Y_train)
acc.append(f1_score(y_test, RF.predict(x_test)))
if np.mean(acc) > accuracy:
accuracy = np.mean(acc)
best_n_estimators = n_estimators
best_split_size = split_size
best_max_features = max_features
print("[NUMBER OF ESTIMATORS]", best_n_estimators)
print("[SPLIT SIZE] ", best_split_size)
print("[MAX FEATURES] ", best_max_features)
print("[ACCURACY] ", accuracy)
model = RandomForestClassifier()
train_data = pd.read_csv("data/train.csv", encoding="latin1")
df = pd.DataFrame(train_data)
labels = train_data['label'].values
x_train = np.array(train_data.drop('label', axis=1))
y_train = labels
x_train = data_preprocess(x_train)
tune(x_train, y_train)
def test_data_preprocess(self):
data = np.ones([5, 3])
result = data_preprocess(data)
self.assertEqual(data.shape[0], result.shape[0])