特征锦囊：怎么批量把特征中的离群点给“安排一下”？

# 导入数据集
import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt

train = pd.read_csv('./data/Group-Image-of-Consumers/train_dataset.csv')
test = pd.read_csv('./data/Group-Image-of-Consumers/test_dataset.csv')
data = pd.concat([train, test], ignore_index=True)
data.head()

# 挑选其中几个变量
feature_list=['当月网购类应用使用次数','当月金融理财类应用使用总次数','当月视频播放类应用使用次数']

# 绘制箱体图
sns.set_style("white")
f, ax = plt.subplots(figsize=(8, 7))
ax.set_xscale("log")
ax = sns.boxplot(data=data[feature_list] , orient="h", palette="Set1")
ax.xaxis.grid(False)
ax.set(ylabel="Feature names")
ax.set(xlabel="Numeric values")
ax.set(title="Numeric Distribution of Features")
sns.despine(trim=True, left=True)

def process(all_data,feature_list):
    #处理离群点
    for col in feature_list:
        ulimit=np.percentile(all_data[col].values, 99.9) #计算一个多维数组的任意百分比分位数
        llimit=np.percentile(all_data[col].values, 0.1)
        all_data.loc[all_data[col]>ulimit,col]=ulimit  # 大于99.9%的直接赋值
        all_data.loc[all_data[col]<llimit,col]=llimit
    return all_data

# 使用函数进行极端值的转换
data_new = process(data,feature_list)

# 绘制箱体图
sns.set_style("white")
f, ax = plt.subplots(figsize=(8, 7))
ax.set_xscale("log")
ax = sns.boxplot(data=data_new[feature_list] , orient="h", palette="Set1")
ax.xaxis.grid(False)
ax.set(ylabel="Feature names")
ax.set(xlabel="Numeric values")
ax.set(title="Numeric Distribution of Features")
sns.despine(trim=True, left=True)