模拟缺失数据1：MCAR

# -*- coding: utf-8 -*-
"""
Created on Mon Sep 14 13:18:53 2020
@author: Dogukan
"""
import numpy as np

def create_mcar_single(df, missing_column, p_missing, random_state=709):
    np.random.seed(random_state)
    indices = [df.sample(n = 1).index[0] for i in range(round(p_missing * df.shape[0]))]
    while len(set(indices)) < round(p_missing * df.shape[0]):
        indices.append(df.sample(n = 1).index[0])
    mcar_column = [1 if i in indices else 0 for i in range(df.shape[0])]
    
    df_new = df.copy()
    for i in range(len(mcar_column)):
        if mcar_column[i] == 1:
            df_new[missing_column][i] = '?'      
    df_new = df_new.replace('?', np.nan)
    return df_new

def create_mcar_mult(df, mising_column, p_missing, random_state):
    df_new = df.copy()
    for i in range(len(mising_column)):
        tmp = create_mcar_single(df, mi