机器学习-KMeans（3D散点图绘制）

读取“客户价值.xlsx”，使用Axes3D绘制反映RFM数据列关系的3D散点图。

import pandas as pd

import numpy as np

from sklearn.cluster import KMeans

df = pd.read_excel('某超市的销售数据.xlsx')

df = df.dropna() # 删除缺失值

df1 = df[['顾客编号', '销售金额', '销售日期','数据采集时间']]

df1['R']=(pd.to_datetime(df1['数据采集时间'])-pd.to_datetime(df1['销售日期'])).values/np.timedelta64(1,'D')

df1=df1[['顾客编号', '销售金额', '销售日期','R']]

df2=df1.groupby('顾客编号').agg({'R':'min','销售金额':'mean'})

df2['F']=df1.groupby(['顾客编号'])['顾客编号'].size()

df2.to_excel('data.xlsx',index=False)

 

datafile=r'data.xlsx'

transformfile=r'tfdata.xls'

data=pd.read_excel(datafile)

data=data[["R",'F','销售金额']]

data=(data-data.mean(axis=0))/(data.std(axis=0))

data.columns=['R','F','M']

data.to_excel(transformfile,index=False)

 

inputfile=r'tfdata.xls'

data = pd.read_excel(inputfile)

iteration=500

kmodel = KMeans(n_clusters = 4,max_iter=iteration)

kmodel.fit(data) # 训练模型

rl=pd.Series(kmodel.labels_).value_counts()

r2=pd.DataFrame(kmodel.cluster_centers_)

r=pd.concat([r2,rl],axis=1)

r.columns=list(data.columns)+[u'聚类数量']

r3 =pd.Series(kmodel.labels_,index=data.index)

r=pd.concat([data,r3],axis=1)

r.columns=list(data.columns)+[u'聚类类别']

r.to_excel('客户价值1.xlsx')

kmodel.cluster_centers_

kmodel.labels_

 

from mpl_toolkits.mplot3d import Axes3D

import matplotlib.pyplot as plt

import matplotlib

matplotlib.rcParams['font.sans-serif']=['SimHei']

matplotlib.rcParams['axes.unicode_minus']=False

fig=plt.figure()

ax=Axes3D(fig)

ax.scatter(r['R'],r['F'],r['M'])

ax.set_xlabel('消费间隔')

ax.set_ylabel('消费频率')

ax.set_zlabel('消费金额')

plt.show