import pandas as pd
import numpy as np
pd.set_option('display.max_columns', None) # 设置显示所有列
df = pd.read_excel('UXFTLB.xls', index_col = 'Id') #读取数据
df.head()

X=1.0*(df - df.mean())/df.std() #数据标准化
X.head()

from yellowbrick.cluster import KElbowVisualizer
from sklearn.cluster import MiniBatchKMeans
visualizer = KElbowVisualizer(MiniBatchKMeans(), k=(1,9))
visualizer.fit(X)
visualizer.poof() # 肘部不明显，分为4类

from yellowbrick.cluster import SilhouetteVisualizer
model = MiniBatchKMeans(4)
visualizer = SilhouetteVisualizer(model)
visualizer.fit(X) 
visualizer.poof()

from sklearn.cluster import KMeans
model = KMeans(n_clusters = 4, n_jobs = 4, max_iter = 500) #分为k类，并发数4，#聚类最大循环次数500
model.fit(X) #开始聚类
KMeans(algorithm='auto', copy_x=True, init='k-means++', max_iter=500,
    n_clusters=4, n_init=10, n_jobs=4, precompute_distances='auto',
    random_state=None, tol=0.0001, verbose=0)
#简单打印结果
r1 = pd.Series(model.labels_).value_counts() #统计各个类别的数目
r2 = pd.DataFrame(model.cluster_centers_) #找出聚类中心
r = pd.concat([r2, r1], axis = 1) #横向连接（0是纵向），得到聚类中心对应的类别下的数目
r.columns = list(X.columns) + ['类别成员数'] #重命名表头
print(r)
R         F         M  类别成员数
0 -0.167313  1.202923  0.187762    308
1  3.551564 -0.256500  0.423271     38
2 -0.138992 -0.681412 -0.802742    351
3 -0.142555 -0.500320  0.855340    243
#详细输出原始数据及其类别
r = pd.concat([df, pd.Series(model.labels_, index = X.index)], axis = 1)  #详细输出每个样本对应的类别
r.columns = list(df.columns) + ['类别'] #重命名表头
r.head()

res = r[r['类别']==3]
res.head(10)