Question

from sklearn import datasets
import pandas as pd
import numpy as np

dt = datasets.load_diabetes()
data = pd.DataFrame(data= np.c_[dt['data'], dt['target']],columns= 
dt['feature_names'] + ['target'] )
data = data.drop('sex', axis = 1)

# mean +- 2sigma
# function to calculate outlier of a variable
def out1(x):
    mu = np.average(x)
    sigma = np.std(x)
    LL = mu - 2*sigma # Lower limit 
    UL = mu + 2*sigma # Upper limit
    out = [1 if (a >= UL) | (a <= LL) else 0 for a in x]
    return(out)

# check #outliers in each variable
print(data.apply(out1).apply(sum))


# Function to Replace outlier with LL / UL

def out_impute(x):
    mu = np.average(x)
    sigma = np.std(x)
    LL = mu - 2*sigma # Lower limit 
    UL = mu + 2*sigma # Upper limit
    xnew = "Enter Code Here"
    return(xnew)

data1 = data.apply(out_impute) # Create new data with inputed values

请有人帮助我如何用上下限代替异常值。

我将异常值定义为> = mu + 2 * sigma和=

谢谢！

Answer 1

使用df.clip

LL = mu - 2*sigma # Lower limit 
UL = mu + 2*sigma # Upper limit
df['data'].clip(LL, UL)

用Python中的下限和上限替换异常值的函数

1 个答案: