群聚編碼問題? - Cupoy

2019/10/17 09:52 下午

機器學習共學討論版

嚴詩豪

觀看數：11

回答數：4

收藏數：0

ml100-3

程式碼:

# 取一個類別型欄位, 與一個數值型欄位, 做群聚編碼

df['Sex'] = df['Sex'].fillna('None')

mean_df = df.groupby(['Sex'])['Age'].mean().reset_index()

mode_df = df.groupby(['Sex'])['Age'].apply(lambda x: x.mode()).reset_index()

median_df = df.groupby(['Sex'])['Age'].median().reset_index()

max_df = df.groupby(['Sex'])['Age'].max().reset_index()

temp = pd.merge(mean_df, mode_df,how='left',on=['Sex'])

temp = pd.merge(temp, median_df,how='left',on=['Sex'])

temp = pd.merge(temp, max_df,how='left',on=['Sex'])

temp.columns = ['Sex','Cabin_age_Mean', 'Cabin_age_Mode', 'Cabin_age_Median', 'Cabin_age_Max','p']

temp

我這樣用出來的欄位怎麼會有六個?