SVM Classifier ▪ 𝐶 = 10−2, 10−1, 1, 10, 102 , 𝛾 = {10−2, 10−1, 1, 10, 102} 20 5. 모델링 적용 및 평가 [3/4] * 매개변수 조합 생략, 최종 매개변수 튜닝 결과 {SVM} SPX SSEC KOSPI GDAXI N225 Score 0.6458 0.4166 0.5208 0.5833 0.5833 Params {'C': 0.01, 'gamma': 0.01} {'C': 0.01, 'gamma': 0.01} {'C': 0.01, 'gamma': 0.01} {'C': 0.01, 'gamma': 0.01} {'C': 0.01, 'gamma': 0.01} {XG boost} SPX SSEC KOSPI GDAXI N225 Score 0.9833 0.9833 0.9500 0.9500 0.9166 Params {'n_estimators':160, 'learning_rate': 0.1, 'reg_alpha':0, 'subsample': 0.76, 'max_depth': 3, 'colsample_bytree': 0.8, 'min_child_weight': 4, 'objective' :'binary:lo gistic', 'scale_pos_weight':1, 'max_delta_step':0, 'gamma' : 0.1} {'n_estimators':160, 'learning_rate': 0.005, 'reg_alpha':0, 'subsample': 0.59, 'max_depth': 3, 'colsample_bytree': 0.4 'min_child_weight': 4, 'objective' : 'binary:logistic', 'scale_pos_weight':1, 'max_delta_step':0, 'gamma' :0.1} {'n_estimators':160, 'learning_rate': 0.1, 'reg_alpha':0, 'subsample': 0.78, 'max_depth': 3, 'colsample_bytree': 0.59, 'min_child_weight': 5, 'objective' :'binary:lo gistic', 'scale_pos_weight':1, 'max_delta_step':0, 'gamma' : 0.1} {'n_estimators':160, 'learning_rate':0.1, 'reg_alpha':0, 'subsample': 0.68, 'max_depth': 3, 'colsample_bytree': 0.51, 'min_child_weight': 2, 'objective' :'binary:lo gistic', 'scale_pos_weight':1, 'max_delta_step':0, 'gamma' :0.1} {'n_estimators':160, 'learning_rate': 0.1, 'reg_alpha':0, 'subsample': 0.44, 'max_depth':4, 'colsample_bytree': 0.64, 'min_child_weight': 0, 'objective' :'binary:lo gistic', 'scale_pos_weight':1, 'max_delta_step':0, 'gamma' : 0.1}