k-means聚类后，利用python计算每个节点到自身聚类中心的欧式距离

最新推荐文章于 2022-12-19 15:35:44 发布

Good_Spark

最新推荐文章于 2022-12-19 15:35:44 发布

阅读量874

点赞数

文章标签： python kmeans 聚类

本文链接：https://blog.csdn.net/qq_31823889/article/details/127792520

版权

利用spark的k-means算子跑完模型后，得到中心节点，计算每个节点到中心节点的距离，用于选择每个聚类中更加典型的节点

import pandas as pd
import sklearn.preprocessing as preproc
import matplotlib.pyplot as plt
import numpy as np

df_center = pd.read_csv('D:\\data\\center\\center.csv', delimiter=',')
df_fen = pd.read_csv('D:\\data\\fenxi\\fenxi.csv', delimiter=',')

clus = df_center['cluster']
itm = 0
for itm in clus:
    df_center22 = df_center.loc[df_center['cluster'] == itm].values
    df_fen22 = df_fen.loc[df_fen['cluster'] == itm].values
    uuuu = df_fen.loc[df_fen['cluster'] == itm].copy()
    uuuu["uu_"+str(itm)] = np.sqrt((np.sum(np.power(df_center22[:, 1:] - df_fen22[:, 1:-1], 2), 1)).astype(float))
    uuuu.to_csv("D:\\data\\center\\resutl_"+str(itm) +".csv", encoding="utf_8_sig")