def data_v4():
X, y_train = load_data()
is_train_obs = X.index.get_level_values('obs_type') == 'train'
X_train, X_test = X[is_train_obs], X[~is_train_obs]
category_cols = [col for col in X.columns if X[col].dtype == 'O']
for col in category_cols:
X_train[col], X_test[col] = encode_with_leave_one_out(
train_col=X_train[col], y=y_train, test_col=X_test[col])
return X_train, y_train, X_test
def data_v4():
X, y_train = load_data()
is_train_obs = X.index.get_level_values('obs_type') == 'train'
X_train, X_test = X[is_train_obs], X[~is_train_obs]
category_cols = [col for col in X.columns if X[col].dtype == 'O']
for col in category_cols:
X_train[col], X_test[col] = encode_with_leave_one_out(
train_col=X_train[col],
y=y_train,
test_col=X_test[col]
)
return X_train, y_train, X_test
def data_v6():
X, y_train = load_data()
is_train_obs = X.index.get_level_values('obs_type') == 'train'
X_train, X_test = X[is_train_obs], X[~is_train_obs]
category_cols = [col for col in X.columns if X[col].dtype == 'O']
for col in category_cols:
X_train[col], X_test[col] = encode_with_leave_one_out(
train_col=X_train[col], y=y_train, test_col=X_test[col])
cols_to_drop = ['T2_V10', 'T2_V7', 'T1_V13', 'T1_V10']
X_train.drop(['T2_V10', 'T2_V7', 'T1_V13', 'T1_V10'], axis=1, inplace=True)
X_test.drop(['T2_V10', 'T2_V7', 'T1_V13', 'T1_V10'], axis=1, inplace=True)
return X_train, y_train, X_test
def data_v6():
X, y_train = load_data()
is_train_obs = X.index.get_level_values('obs_type') == 'train'
X_train, X_test = X[is_train_obs], X[~is_train_obs]
category_cols = [col for col in X.columns if X[col].dtype == 'O']
for col in category_cols:
X_train[col], X_test[col] = encode_with_leave_one_out(
train_col=X_train[col],
y=y_train,
test_col=X_test[col]
)
cols_to_drop = ['T2_V10', 'T2_V7', 'T1_V13', 'T1_V10']
X_train.drop(['T2_V10', 'T2_V7', 'T1_V13', 'T1_V10'], axis=1, inplace=True)
X_test.drop(['T2_V10', 'T2_V7', 'T1_V13', 'T1_V10'], axis=1, inplace=True)
return X_train, y_train, X_test
