인프런 커뮤니티 질문&답변

CV를 사용하였는데 하기와 같은 오류가 등장했습니다 :

--> 98             raise InvalidParameterError(
     99                 f"The {param_name!r} parameter of {caller_name} must be"
    100                 f" {constraints_str}. Got {param_val!r} instead."

InvalidParameterError: The 'scoring' parameter of cross_val_score must be a str among {'f1', 'jaccard_micro', 'positive_likelihood_ratio', 'adjusted_rand_score', 'jaccard_weighted', 'homogeneity_score', 'average_precision', 'precision_weighted', 'rand_score', 'roc_auc_ovr', 'roc_auc_ovr_weighted', 'precision', 'explained_variance', 'jaccard_macro', 'recall_macro', 'f1_macro', 'normalized_mutual_info_score', 'precision_samples', 'neg_root_mean_squared_log_error', 'r2', 'neg_negative_likelihood_ratio', 'precision_micro', 'neg_max_error', 'mutual_info_score', 'precision_macro', 'f1_micro', 'v_measure_score', 'completeness_score', 'neg_mean_squared_error', 'accuracy', 'neg_brier_score', 'recall_samples', 'jaccard_samples', 'neg_root_mean_squared_error', 'neg_mean_absolute_percentage_error', 'jaccard', 'f1_samples', 'matthews_corrcoef', 'neg_median_absolute_error', 'neg_mean_gamma_deviance', 'recall_micro', 'neg_mean_absolute_error', 'neg_log_loss', 'roc_auc_ovo_weighted', 'd2_absolute_error_score', 'roc_auc', 'adjusted_mutual_info_score', 'recall', 'recall_weighted', 'balanced_accuracy', 'f1_weighted', 'top_k_accuracy', 'roc_auc_ovo', 'neg_mean_squared_log_error', 'fowlkes_mallows_score', 'neg_mean_poisson_deviance'}, a callable or None. Got 'f1-macro' instead.

제가 작성한 코드도 함께 공유 드립니다 :

import pandas as pd
train=pd.read_csv('train.csv')
test=pd.read_csv('test.csv')

train=train.drop('ID', axis=1)
test_id=test.pop('ID')

print(train.shape, test.shape)

print(train.head())
print(test.head())

print(train.info())
print(test.info())
print(train.isnull().sum())
print(test.isnull().sum())

y=train.pop('Segmentation')

print(y.info(), y.shape)

y=y-1

int_cols=train.select_dtypes(exclude='object').columns
train[int_cols].corr()

cat_cols=train.select_dtypes(include='object').columns
print(train[cat_cols].describe(include='object'))
print(test[cat_cols].describe(include='object'))

for i in cat_cols:
  train[i]=train[i].astype('object')
  test[i]=test[i].astype('object')

for i in cat_cols:
  print(train[i].value_counts())
  print(test[i].value_counts())


from sklearn.preprocessing import RobustScaler
scaler=RobustScaler()

for i in int_cols :
  train[i]=scaler.fit_transform(train[[i]])
  test[i]=scaler.transform(test[[i]])


from sklearn.preprocessing import LabelEncoder
le=LabelEncoder()

for i in cat_cols:
  train[i]=le.fit_transform(train[i])
  test[i]=le.transform(test[i])

print(train.head())
print(test.head())

from sklearn.model_selection import train_test_split
X_tr, X_val, y_tr, y_val=train_test_split(train, y, test_size=0.2, random_state=2025)

print(X_tr.shape, X_val.shape, y_tr.shape, y_val.shape)


from sklearn.ensemble import RandomForestClassifier
rf=RandomForestClassifier(n_estimators=50, max_depth=7, random_state=2025)
rf.fit(X_tr, y_tr)
y_pred_rf=rf.predict(X_val)


from lightgbm import LGBMClassifier
lgbm=LGBMClassifier(random_state=2025)
lgbm.fit(X_tr, y_tr)
y_pred_lgbm=lgbm.predict(X_val)

from xgboost import XGBClassifier
xgb=XGBClassifier(random_state=2025)
xgb.fit(X_tr, y_tr)
y_pred_xgb=xgb.predict(X_val)
y_pred_xgb=y_pred_xgb+1


from sklearn.metrics import f1_score
print(f1_score(y_val, y_pred_rf, average='macro'))
print(f1_score(y_val, y_pred_lgbm, average='macro'))
print(f1_score(y_val, y_pred_xgb, average='macro'))

from sklearn.model_selection import cross_val_score
scores=cross_val_score(rf, train, target, scoring='f1-macro', cv=5)
print(scores)
print(scores.mean())