결측치 제거
df <- df[complete.cases(df), ]
y_train_pd의 'pred-true' 변수의 outlier를 제거
#remove outlier
Q1 = y_train_pd['pred-true'].quantile(0.25)
Q3 = y_train_pd['pred-true'].quantile(0.75)
IQR = Q3 - Q1 #IQR is interquartile range.
filter = (y_train_pd['pred-true'] >= Q1 - 1.5 * IQR) & (y_train_pd['pred-true'] <= Q3 + 1.5 *IQR)
ro_list = y_train_pd.loc[filter]
print(y_train_pd['pred-true'].describe())
print(ro_list['pred-true'].describe())
