python实现组合优化

最新推荐文章于 2024-11-21 07:30:52 发布

沸点数据

最新推荐文章于 2024-11-21 07:30:52 发布

阅读量2.5k

点赞数

分类专栏：统计学组合优化文章标签：数据挖掘 python

本文链接：https://blog.csdn.net/weixin_38239039/article/details/105732360

版权

统计学同时被 2 个专栏收录

2 篇文章

订阅专栏

组合优化

1 篇文章

订阅专栏

1、多个变量组合与单一目标，实现组合优化：

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import itertools
%matplotlib inline

#  生成结果字典，转化为DataFrame
result={
    "gender":[],
    "age":[],
    "job":[],
    "job_time":[],
    "size":[],
    "mean":[]
}
metricx="salsry"
data=pd.DataFrame(result)

# 指标之间组合交叉分析
def group_combine(data,dimension,metric):
    data_agg=data.groupby(data[list(dimension)])["metric"].agg([np.size,np.mean]).set_index()
    return data_agg.to_dict(orient="records")


# 将结果追加到dataframe中
def append_data(agg_dict):
    for line in agg_dict:
        for element in datframe_elements:
            result_dict[element].append(line[element]) if element in line else result_dict[element].append('')


for number in range(len(elements)):
    print number
    print list(combinations(elements, number+1))
    for combination in list(combinations(elements, number+1)): 
        print list(combination)
        print metric
        agg_dict = group_aggregation(data = df1, dimension = list(combination), metric = metric)
        append_data(agg_dict)