目录
1.关于聚类
import pandas as pd
import numpy as np
from pandas import Series,DataFrame
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline
聚类的基本使用
# 手动随机生成随机点来做实验
from sklearn.datasets import make_blobs
X, y = make_blobs(n_samples=150, n_features=2, centers=3 ,cluster_std=1.5, random_state=2)
X,y
(array([[ 1.53194956, -0.36022153],
[ -1.56430585, -9.59730336],
[ -1.08878137, -0.53673972],
[ 2.43442247, -0.15599663],
[ 0.48019529, -12.99688015],
[ 0.57898032, -2.06887799],
[ -0.72443515, -7.44202457],
[ -1.85225071, -3.98632318],
[ 2.54279316, -1.7870558 ],
[ -0.52731615, -10.74779592],
[ -2.29661532, -11.6406339 ],
[ 0.92141506, -9.98499137],
[ -5.35083116, -0.65431189],
[ -0.88989127, 0.11369336],
[ -0.59631184, -2.29097658],
[ -1.50195028, -5.4011869 ],
[ -0.76364841, -4.6539681 ],
[ 2.19201955, 0.60036835],
[ 7.15628849, -0.06187083],
[ 4.81890691, -3.50331202],
[ 0.65087822, -4.39797054],
[ -3.24158992, -4.41559955],
[ -1.13496627, -1.67121333],
[ 0.03606565, -2.04003449],
[ -2.01792323, -2.58566719],
[ 1.355409 , -0.54741367],
[ -3.12791644, -4.06556581],
[ -5.26927614, -9.6186543 ],
[ -0.02442698, -1.33977954],
[ -2.86703029, -10.84498679],
[ 2.29764685, -2.92418801],
[ -1.18679697, -1.80057881],
[ -0.73325486, -1.93333585],
[ -0.52577983, -11.34940749],
[ -0.50461407, -3.93251527],
[ 1.90846569, -0.6583068 ],
[ -2.06104996, -0.17628645],
[ -0.47151448, -10.37571491],
[ 1.26386427, -0.46380574],
[ -0.36309079, -9.40951948],
[ -0.53887254, -0.6449586 ],
[ -2.23212091, -8.718881 ],
[ -4.16374507, -3.50826293],
[ -0.70730261, -8.6320622 ],
[ 1.18048503, -0.15879893],
[ -1.41098559, -4.66354671],
[ -1.90907668, -9.67996871],
[ 2.03754653, -0.24742774],
[ -0.97378999, -7.371431 ],
[ 2.99659881, -0.83960814],
[ -3.4119278 , -9.71171816],
[ 0.10647516, -2.83784632],
[ 3.68213884, -1.93707213],
[ -1.88594036, -11.55825336],
[ 0.46218028, -8.90235829],
[ -0.04304745, -7.60915598],
[ -1.78833491, -9.83575141],
[ -0.95592795, -12.26939394],
[ -1.32676236, -4.41753005],
[ 0.22044687, -10.05311414],
[ -4.0709444 , -4.40679626],
[ -4.62768987, -2.99134472],
[ -4.36824992, -2.89757148],
[ -4.08223794, -6.88469836],
[ 0.77102877, 0.95860323],
[ -0.05463537, -2.68689003],
[ 0.44653092, -2.60752136],
[ -0.49365731, -8.378556 ],
[ 0.65278373, -5.68997024],
[ -3.32769271, -1.54225156],
[ -0.36011954, -2.18001056],
[ 2.2374372 , -0.3476192 ],
[ 1.5880298 , -1.7654783 ],
[ -2.31262163, -4.92277723],
[ -0.28638281, -2.50409338],
[ -0.62985746, -7.56390652],
[ 1.30709149, -4.99949807],
[ -3.88704121, -7.92023943],
[ 0.31190778, -0.52199607],
[ -0.70822817, -2.35468348],
[ -2.44971637, -2.95465548],
[ -1.71601202, -3.85030346],
[ -2.06618377, -5.41830673],
[ -1.46459731, -2.39530216],
[ -2.52380489, -9.34991004],
[ -1.83223015, -2.56988374],
[ -1.02782509, -3.59652323],
[ -2.51078608, -3.92019727],
[ -2.63990045, -3.03337678],
[ -1.73623162, -5.60353306],
[ -0.89524628, -10.96464394],
[ -5.15424798, -3.30552368],
[ -3.3851438 , -4.1251994 ],
[ 2.67007966, -1.70491528],
[ 2.12119683, -2.78419362],
[ -2.25997736, -8.21779094],
[ 1.74015978, -1.10379588],
[ 3.29089003, -4.27232081],
[ -1.54379575, -5.85414392],
[ -1.75036425, -8.32495776],
[ -1.33945732, -8.99247021],
[ -2.92821038, -7.10474478],
[ -1.00719928, -1.93003946],
[ 0.29073017, -3.17563261],
[ -1.28008731, -8.66794651],
[ 1.54082983, -0.1324291 ],
[ -1.84360609, -9.59318151],
[ -0.597949 , -0.40605237],
[ -2.23658448, -11.26289379],
[ -3.19324464, -4.3727003 ],
[ -1.21779287, -11.15836353],
[ -2.86763721, -4.67181627],
[ 0.82161761, -2.04081344],
[ -0.45292089, -6.04316334],
[ -0.709394 , -9.80717827],
[ -4.93225332, -9.31238561],
[ -0.23742255, -12.53167518],
[ -2.40190838, -9.46793749],
[ 2.65696448, -3.94092874],
[ 0.10261618, 0.4306987 ],
[ 0.77075118, -7.65464691],
[ -1.97310998, -8.95514262],
[ -1.23044866, -0.02408431],
[ -0.83889419, 1.41316281],
[ 1.89552328, -1.28806291],
[ 3.74624864, -0.63251734],
[ 0.50567512, -2.13390391],
[ -3.56899486, -6.43169397],
[ -1.30528349, -4.3866171 ],
[ -2.97980187, -8.83183653],
[ -1.85237668, -9.38174185],
[ 3.60596784, -1.96480346],
[ -1.59595363, -3.6022414 ],
[ -1.32254393, -3.49370015],
[ -4.34058653, -9.41209208],
[ -0.05036661, -4.47612317],
[ -1.60404567, -4.79404957],
[ -2.52020719, -2.82511188],
[ -2.5972638 , -9.71612662],
[ -0.82276679, -3.89556977],
[ 5.91286766, 0.16273983],
[ -1.85513922, -5.54901873],
[ 0.7183647 , 0.23622995],
[ -1.6836874 , -6.13442518],
[ -0.0856312 , -2.16867404],
[ -1.3087977 , -7.71897353],
[ -2.42206812, -2.94401336],
[ 0.50787945, -0.65781509],
[ -3.18469257, -3.55607882],
[ -1.50676754, -3.15467085]]),
array([1, 0, 2, 1, 0, 1, 0, 2, 1, 0, 0, 0, 2, 1, 2, 2, 2, 1, 1, 1, 1, 2,
1, 2, 2, 1, 2, 0, 2, 0, 1, 1, 1, 0, 2, 1, 2, 0, 1, 0, 1, 0, 2, 0,
1, 2, 0, 1, 0, 1, 0, 2, 1, 0, 0, 0, 0, 0, 1, 0, 2, 2, 2, 0, 1, 1,
1, 0, 2, 2, 1, 1, 1, 2, 2, 0, 2, 0, 1, 1, 2, 2, 2, 2, 0, 2, 2, 2,
2, 0, 0, 2, 2, 1, 1, 0, 1, 1, 2, 0, 0, 0, 2, 1, 0, 1, 0, 1, 0, 2,
0, 2, 1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 1, 1, 1, 1, 2, 2, 0, 0, 1,
2, 2, 0, 2, 2, 2, 0, 2, 1, 2, 1, 0, 1, 0, 2, 1, 2, 2]))
sns.set()
X.shape
(150, 2)
y.shape
(150,)
plt.scatter(X[:,0],X[:,1],c=y,cmap=plt.cm.Accent_r)
<matplotlib.collections.PathCollection at 0x1a3478924f0>
?
?
# 导入kmeans
from sklearn.cluster import KMeans
# 1. 实例化Kmeans对象
# kmeans下的重要参数： n_clusters 默认是8 即8个随机点（随机种子） 分成8个簇 类型int
# k值怎么挑选： 看业务需求 根据业务需要
km = KMeans(n_clusters=3)
# km.fit_predict() 在kmeans这种聚类算法中 sklearn为我们集成了两种方式，上下两种都可以使用
#km.fit()
#km.predict()
# 2.fit_predict 无监督学习直接预测 训练预测二合一
y_ = km.fit_predict(X)
y_
array([2, 1, 2, 2, 1, 2, 1, 0, 2, 1, 1, 1, 0, 2, 0, 0, 0, 2, 2, 2, 0, 0,
0, 2, 0, 2, 0, 1, 2, 1, 2, 0, 0, 1, 0, 2, 0, 1, 2, 1, 2, 1, 0, 1,
2, 0, 1, 2, 1, 2, 1, 0, 2, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 1, 2, 0,
2, 1, 0, 0, 0, 2, 2, 0, 0, 1, 0, 1, 2, 0, 0, 0, 0, 0, 1, 0, 0, 0,
0, 0, 1, 0, 0, 2, 2, 1, 2, 2, 0, 1, 1, 1, 0, 0, 1, 2, 1, 2, 1, 0,
1, 0, 2, 0, 1, 1, 1, 1, 2, 2, 1, 1, 2, 2, 2, 2, 2, 0, 0, 1, 1, 2,
0, 0, 1, 0, 0, 0, 1, 0, 2, 0, 2, 0, 2, 1, 0, 2, 0, 0])
# 获取最后的聚类中心
cluster_centers = km.cluster_centers_
cluster_centers
array([[-1.80632868, -3.67173199],
[-1.62473796, -9.4792349 ],
[ 1.58887503, -1.06495221]])
plt.figure(figsize=(12,4))
ax1 = plt.subplot(1,2,1)
ax2 = plt.subplot(1,2,2)
ax1.scatter(X[:,0],X[:,1],c=y,cmap=plt.cm.cool)
ax2.scatter(X[:,0],X[:,1],c=y_,cmap=plt.cm.rainbow_r)
ax2.scatter(cluster_centers[:,0],cluster_centers[:,1],color ='red',marker = '*',s=300)
ax1.set_title('True')
ax2.set_title('Kmeans')
plt.show()
?
?
球队综合实力聚类分析
data = pd.read_csv('AsiaZoo.txt',header=None)
data
| 0 | 1 | 2 | 3 | |
|---|---|---|---|---|
| 0 | 中国 | 50 | 50 | 9 |
| 1 | 日本 | 28 | 9 | 4 |
| 2 | 韩国 | 17 | 15 | 3 |
| 3 | 伊朗 | 25 | 40 | 5 |
| 4 | 沙特 | 28 | 40 | 2 |
| 5 | 伊拉克 | 50 | 50 | 1 |
| 6 | 卡塔尔 | 50 | 40 | 9 |
| 7 | 阿联酋 | 50 | 40 | 9 |
| 8 | 乌兹别克斯坦 | 40 | 40 | 5 |
| 9 | 泰国 | 50 | 50 | 9 |
| 10 | 越南 | 50 | 50 | 5 |
| 11 | 阿曼 | 50 | 50 | 9 |
| 12 | 巴林 | 40 | 40 | 9 |
| 13 | 朝鲜 | 40 | 32 | 17 |
| 14 | 印尼 | 50 | 50 | 9 |
data.columns = ['国家','2006年世界杯','2010年世界杯','2007年亚洲杯']
data
| 国家 | 2006年世界杯 | 2010年世界杯 | 2007年亚洲杯 | |
|---|---|---|---|---|
| 0 | 中国 | 50 | 50 | 9 |
| 1 | 日本 | 28 | 9 | 4 |
| 2 | 韩国 | 17 | 15 | 3 |
| 3 | 伊朗 | 25 | 40 | 5 |
| 4 | 沙特 | 28 | 40 | 2 |
| 5 | 伊拉克 | 50 | 50 | 1 |
| 6 | 卡塔尔 | 50 | 40 | 9 |
| 7 | 阿联酋 | 50 | 40 | 9 |
| 8 | 乌兹别克斯坦 | 40 | 40 | 5 |
| 9 | 泰国 | 50 | 50 | 9 |
| 10 | 越南 | 50 | 50 | 5 |
| 11 | 阿曼 | 50 | 50 | 9 |
| 12 | 巴林 | 40 | 40 | 9 |
| 13 | 朝鲜 | 40 | 32 | 17 |
| 14 | 印尼 | 50 | 50 | 9 |
from mpl_toolkits.mplot3d import Axes3D
# 设置中文以及seaborn
sns.set_style(style='white')
plt.rcParams['font.sans-serif'] = ['SimHei']
plt.rcParams['axes.unicode_minus'] = False
# 下面进行画图 用3D绘制 （3个特征）
plt.figure(figsize=(10,6))
ax = plt.subplot(projection='3d')
ax.scatter3D(data['2006年世界杯'],data['2010年世界杯'],data['2007年亚洲杯'],s = 200,cmap = plt.cm.rainbow)
ax.set_xlabel('2006年世界杯')
ax.set_ylabel('2010年世界杯')
ax.set_zlabel('2007年亚洲杯')
plt.show()
?
?
# 构造X的数据 解决‘国家’
X = data.iloc[:,1:]
X
| 2006年世界杯 | 2010年世界杯 | 2007年亚洲杯 | |
|---|---|---|---|
| 0 | 50 | 50 | 9 |
| 1 | 28 | 9 | 4 |
| 2 | 17 | 15 | 3 |
| 3 | 25 | 40 | 5 |
| 4 | 28 | 40 | 2 |
| 5 | 50 | 50 | 1 |
| 6 | 50 | 40 | 9 |
| 7 | 50 | 40 | 9 |
| 8 | 40 | 40 | 5 |
| 9 | 50 | 50 | 9 |
| 10 | 50 | 50 | 5 |
| 11 | 50 | 50 | 9 |
| 12 | 40 | 40 | 9 |
| 13 | 40 | 32 | 17 |
| 14 | 50 | 50 | 9 |
# 1.实例化kmeans
km = KMeans(n_clusters=3)
# 2.预测 fit_predict
y_ = km.fit_predict(X)
y_
array([0, 1, 1, 2, 2, 0, 0, 0, 2, 0, 0, 0, 2, 2, 0])
# 下面进行画图 用3D绘制 （3个特征）
plt.figure(figsize=(10,6))
ax = plt.subplot(projection='3d')
ax.scatter3D(data['2006年世界杯'],data['2010年世界杯'],data['2007年亚洲杯'],s = 200,c = y_,cmap = plt.cm.rainbow)
ax.set_xlabel('2006年世界杯')
ax.set_ylabel('2010年世界杯')
ax.set_zlabel('2007年亚洲杯')
plt.show()
# 在元数据集中做增加column的操作 使得聚类结果展示出来
data['簇类'] = y_
data
| 国家 | 2006年世界杯 | 2010年世界杯 | 2007年亚洲杯 | 簇类 | |
|---|---|---|---|---|---|
| 0 | 中国 | 50 | 50 | 9 | 0 |
| 1 | 日本 | 28 | 9 | 4 | 1 |
| 2 | 韩国 | 17 | 15 | 3 | 1 |
| 3 | 伊朗 | 25 | 40 | 5 | 2 |
| 4 | 沙特 | 28 | 40 | 2 | 2 |
| 5 | 伊拉克 | 50 | 50 | 1 | 0 |
| 6 | 卡塔尔 | 50 | 40 | 9 | 0 |
| 7 | 阿联酋 | 50 | 40 | 9 | 0 |
| 8 | 乌兹别克斯坦 | 40 | 40 | 5 | 2 |
| 9 | 泰国 | 50 | 50 | 9 | 0 |
| 10 | 越南 | 50 | 50 | 5 | 0 |
| 11 | 阿曼 | 50 | 50 | 9 | 0 |
| 12 | 巴林 | 40 | 40 | 9 | 2 |
| 13 | 朝鲜 | 40 | 32 | 17 | 2 |
| 14 | 印尼 | 50 | 50 | 9 | 0 |
data.groupby('簇类').groups
{0: [0, 5, 6, 7, 9, 10, 11, 14], 1: [1, 2], 2: [3, 4, 8, 12, 13]}
data.groupby('簇类').groups.items()
dict_items([(0, Int64Index([0, 5, 6, 7, 9, 10, 11, 14], dtype='int64')), (1, Int64Index([1, 2], dtype='int64')), (2, Int64Index([3, 4, 8, 12, 13], dtype='int64'))])
# 复习一下 for循环输出不同的国家和梯队
for _,indexes in data.groupby('簇类').groups.items():
countries = data.loc[indexes,'国家']
for country in countries:
print(country, end=' ')
print()
中国 伊拉克 卡塔尔 阿联酋 泰国 越南 阿曼 印尼
日本 韩国
伊朗 沙特 乌兹别克斯坦 巴林 朝鲜
2.kmeans中常见的错误
import pandas as pd
import numpy as np
from pandas import Series,DataFrame
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline
from sklearn.cluster import KMeans
from sklearn.datasets import make_blobs
a. k值的不合理不合适
# k值的不合理不合适
X ,y = make_blobs(n_samples=150, n_features=2, centers=3, random_state=2, cluster_std=2)
sns.set()
# 将画图函数进行封装 画图时传入X y即可快速画图
def show_scatter(X,y):
plt.scatter(X[:,0],X[:,1],c=y,cmap=plt.cm.rainbow)
plt.show()
# 用这个封装的函数进行绘制
show_scatter(X,y)
?
?
# 实例化一个kmeans对象
km = KMeans(n_clusters=2)
y_ = km.fit_predict(X)
# 封装一个函数 展现y 和y_ 的对比 【 true 和 predict】
def show_predict(X,y,y_):
# 设置画布比例
plt.figure(figsize=(12,4))
# 设置子图 121代表位置在左边
plt.subplot(121)
plt.scatter(X[:,0],X[:,1],c=y,s=100,cmap=plt.cm.rainbow)
plt.title('True')
# 设置子图 122代表位置在右边
plt.subplot(122)
plt.scatter(X[:,0],X[:,1],c=y_,s=100,cmap=plt.cm.cool)
plt.title('Kmeans prediction')
plt.show()
# 调用函数进行图像绘制
show_predict(X,y,y_)
?
?
km = KMeans(n_clusters=7)
y_ = km.fit_predict(X)
show_predict(X,y,y_)
?
b. 数据存在偏差 （必然存在）
# 去模仿数据存在偏差的状况（极端情况）
trans = [[0.6,-0.6],[-0.4,0.8]]
X1 = np.dot(X, trans)
X1
array([[ 1.04655426, -1.06619879],
[ 2.85894079, -6.71330586],
[ -0.7186756 , 0.88485489],
[ 1.65961265, -1.57033723],
[ 6.30764932, -11.97512202],
[ 1.19546232, -2.12639029],
[ 2.38135532, -5.08623838],
[ 0.51032678, -2.18392533],
[ 2.61620742, -3.39683023],
[ 4.30212858, -8.77008969],
[ 3.3628695 , -8.30701087],
[ 5.05428445, -9.11541647],
[ -4.06561027, 4.1690844 ],
[ -1.14367301, 1.37678308],
[ 0.61089302, -1.38030672],
[ 1.5451611 , -3.97335364],
[ 1.7372859 , -3.76696174],
[ 1.06229566, -0.56962559],
[ 5.38690504, -5.2474292 ],
[ 5.35243507, -7.04839453],
[ 2.49516333, -4.668274 ],
[ -0.3721972 , -1.53034875],
[ -0.38778277, -0.33105738],
[ 0.98295923, -1.61853714],
[ -0.36922777, -0.55735425],
[ 1.00515763, -1.12463797],
[ -0.46794307, -1.24791822],
[ -0.09364828, -3.77210397],
[ 0.56109582, -0.82320444],
[ 2.4821924 , -7.00198864],
[ 3.02656089, -4.41365421],
[ -0.3602524 , -0.42758268],
[ 0.07338504, -0.93202387],
[ 4.62421714, -9.41303776],
[ 1.55973854, -3.20463953],
[ 1.50674598, -1.68536932],
[ -1.68873222, 2.04715326],
[ 4.14832672, -8.41784462],
[ 0.88733094, -0.96222038],
[ 3.71976143, -7.4739751 ],
[ -0.45824364, 0.28673935],
[ 1.85619681, -5.24206996],
[ -1.59383419, 0.1752011 ],
[ 3.0297481 , -6.36931789],
[ 0.65795725, -0.5701764 ],
[ 1.22452476, -3.25930919],
[ 2.62721231, -6.52566558],
[ 1.39087515, -1.35036299],
[ 2.14422155, -4.8114547 ],
[ 2.47394653, -2.74926392],
[ 1.44186445, -5.35725076],
[ 1.46478647, -2.52586403],
[ 3.60769267, -4.46832419],
[ 3.64747318, -8.54767826],
[ 4.10949232, -7.59322003],
[ 3.01560224, -5.80962205],
[ 2.80688984, -6.78842721],
[ 4.77075809, -10.05023814],
[ 0.92348261, -3.10702502],
[ 4.52984205, -8.62730621],
[ -1.04037586, -0.857475 ],
[ -2.24067973, 1.09773636],
[ -2.08314016, 0.99020918],
[ -0.60212755, -1.80551487],
[ -0.26555557, 0.94928424],
[ 1.01817619, -2.27871059],
[ 1.37677927, -2.59498371],
[ 3.06546103, -6.26982751],
[ 3.42296609, -6.00517641],
[ -1.97353169, 1.60343801],
[ 0.50345313, -1.49365181],
[ 1.60422314, -1.61714642],
[ 1.84088873, -2.61000354],
[ 0.6414722 , -2.81451291],
[ 0.97249854, -1.85557453],
[ 2.52202118, -5.29190794],
[ 3.57816048, -5.79211897],
[ 0.1063184 , -3.06624939],
[ 0.15680059, -0.26272488],
[ 0.31812512, -1.40148269],
[ -0.51786853, -0.60550725],
[ 0.54677387, -2.14782857],
[ 1.10290489, -3.540228 ],
[ -0.02809505, -0.79695896],
[ 1.95939845, -5.68182042],
[ -0.22909115, -0.68907304],
[ 0.96197396, -2.42767921],
[ -0.05176867, -1.58656273],
[ -0.6280311 , -0.53732937],
[ 0.59138935, -2.3157436 ],
[ 4.12343675, -8.70705014],
[ -2.49436411, 1.18385863],
[ -0.64192038, -1.10574549],
[ 2.67422834, -3.41104287],
[ 2.8107372 , -4.12316684],
[ 1.56666362, -4.68528874],
[ 1.60969543, -2.02591295],
[ 4.54015959, -6.64625707],
[ 1.7532618 , -4.42303141],
[ 2.03150974, -5.2072905 ],
[ 2.71624194, -6.24802934],
[ 0.43845259, -2.96345309],
[ 0.08968327, -0.66659717],
[ 1.55513133, -3.07632843],
[ 2.59065863, -5.94936671],
[ 0.93216919, -0.83032442],
[ 2.63330227, -6.48546903],
[ -0.63292147, 0.58883382],
[ 3.20943279, -7.95211276],
[ -0.35640058, -1.52326577],
[ 3.9687166 , -8.6556471 ],
[ 0.06361389, -2.10280876],
[ 1.37460439, -2.29056461],
[ 1.85250741, -3.81133115],
[ 3.65480355, -7.62110191],
[ 0.01262668, -3.71503562],
[ 5.48544573, -10.90480912],
[ 2.11986363, -5.90523358],
[ 3.85627671, -5.78563176],
[ -0.51873656, 0.92091616],
[ 3.69090297, -6.50918461],
[ 2.18941176, -5.70129111],
[ -1.34263749, 1.50226615],
[ -1.79592572, 2.72208617],
[ 1.83226198, -2.34675524],
[ 2.96321796, -3.12808692],
[ 1.17149864, -2.13710711],
[ 0.44112921, -3.41892551],
[ 1.16139064, -3.04847928],
[ 1.318295 , -4.7644111 ],
[ 2.51351799, -6.25291692],
[ 3.56154592, -4.43696748],
[ 0.51052083, -1.97927577],
[ 0.67135991, -2.08222618],
[ 0.53913689, -4.29472262],
[ 2.21306072, -4.14788593],
[ 1.13967822, -3.24406418],
[ -0.6433511 , -0.41093476],
[ 2.09594684, -6.01368432],
[ 1.28551209, -2.91070882],
[ 4.27237602, -4.01310782],
[ 1.34145359, -3.84848977],
[ 0.07757892, 0.220884 ],
[ 0.91656719, -2.9240639 ],
[ 0.71699766, -1.70115019],
[ 2.0615714 , -4.91416057],
[ -0.50142572, -0.61627426],
[ 0.38601474, -0.56437584],
[ -0.78509038, -0.65904451],
[ 0.3431654 , -1.57321605]])
show_scatter(X1,y)
show_scatter(X,y)
# 首先实例化
km2 =KMeans(n_clusters=3)
# 预测：根据存在偏差的数据进行预测
y2_ = km2.fit_predict(X1)
show_predict(X , y ,y2_)
?
c. 标准偏差不相同cluster_std
元 万元
8000 0.8
X , y = make_blobs(n_samples=150, n_features=2, random_state=2, cluster_std=[1,2,4])
show_scatter(X,y)
?
km3 = KMeans(n_clusters=3)
y3_ = km3.fit_predict(X)
show_predict(X,y,y3_)
from sklearn.preprocessing import StandardScaler
ss_X = StandardScaler().fit_transform(X)
ss_X.std(axis=0)
array([1., 1.])
d. 样本数量不同
X,y = make_blobs(n_samples=1500, n_features=2, centers=3, random_state=5)
# 故意按照100 35 15 取数据 以使得样本数量不同
# 造X的假数据
XA = X[y==0][:100]
XB = X[y==1][:35]
XC = X[y==2][:15]
XX = np.concatenate((XA,XB,XC))
# 造y的数据
y = np.array([0]*100 + [1]*35 + [2]*15)
XX.shape
(150, 2)
y.shape
(150,)
show_scatter(XX,y)
?
km4 = KMeans(n_clusters=3)
y4_ = km4.fit_predict(XX)
show_predict(XX, y, y4_)
?
?
e. 使用轮廓系数来判断聚类的效果
X ,y = make_blobs(n_samples=150, n_features=2, centers=3, random_state=3,cluster_std=3)
show_scatter(X,y)
?
from sklearn.metrics import silhouette_score
kmeans = KMeans(n_clusters=3)
y_ = kmeans.fit_predict(X)
# 使用轮廓系数进行评分
silhouette_score(X,y_)
0.39860514202079084
# 定义一个函数进行绘图，以发现不同的轮廓系数的图像和得分
def show_clusters_edge(kmeans, X):
# 画底图
xmin, xmax = X[:,0].min(), X[:,0].max()
ymin, ymax = X[:,1].min(), X[:,1].max()
x = np.linspace(xmin, xmax, 200)
y = np.linspace(ymin, ymax,200)
xx , yy = np.meshgrid(x, y)
kmeans.fit(X)
X_test = np.c_[xx.ravel(), yy.ravel()]
y_ = kmeans.predict(X)
y1_ = kmeans.predict(X_test)
plt.scatter(X_test[:,0],X_test[:,1], c=y1_,s=100, cmap=plt.cm.Accent)
plt.scatter(X[:,0],X[:,1],c=y_,s= 100,cmap=plt.cm.cool)
plt.title('silhouette_score：%.4f'%(silhouette_score(X,y_)))
plt.show()
kmeans = KMeans(n_clusters=3)
show_clusters_edge(kmeans,X)
D:\software\anaconda\lib\site-packages\matplotlib\backends\backend_agg.py:240: RuntimeWarning: Glyph 65306 missing from current font.
font.set_text(s, 0.0, flags=flags)
D:\software\anaconda\lib\site-packages\matplotlib\backends\backend_agg.py:203: RuntimeWarning: Glyph 65306 missing from current font.
font.set_text(s, 0, flags=flags)
kmeans = KMeans(n_clusters=4)
show_clusters_edge(kmeans,X)
D:\software\anaconda\lib\site-packages\matplotlib\backends\backend_agg.py:240: RuntimeWarning: Glyph 65306 missing from current font.
font.set_text(s, 0.0, flags=flags)
D:\software\anaconda\lib\site-packages\matplotlib\backends\backend_agg.py:203: RuntimeWarning: Glyph 65306 missing from current font.
font.set_text(s, 0, flags=flags)
kmeans = KMeans(n_clusters=5)
show_clusters_edge(kmeans,X)
D:\software\anaconda\lib\site-packages\matplotlib\backends\backend_agg.py:240: RuntimeWarning: Glyph 65306 missing from current font.
font.set_text(s, 0.0, flags=flags)
D:\software\anaconda\lib\site-packages\matplotlib\backends\backend_agg.py:203: RuntimeWarning: Glyph 65306 missing from current font.
font.set_text(s, 0, flags=flags)
kmeans = KMeans(n_clusters=6)
show_clusters_edge(kmeans,X)
D:\software\anaconda\lib\site-packages\matplotlib\backends\backend_agg.py:240: RuntimeWarning: Glyph 65306 missing from current font.
font.set_text(s, 0.0, flags=flags)
D:\software\anaconda\lib\site-packages\matplotlib\backends\backend_agg.py:203: RuntimeWarning: Glyph 65306 missing from current font.
font.set_text(s, 0, flags=flags)
kmeans = KMeans(n_clusters=7)
show_clusters_edge(kmeans,X)
D:\software\anaconda\lib\site-packages\matplotlib\backends\backend_agg.py:240: RuntimeWarning: Glyph 65306 missing from current font.
font.set_text(s, 0.0, flags=flags)
D:\software\anaconda\lib\site-packages\matplotlib\backends\backend_agg.py:203: RuntimeWarning: Glyph 65306 missing from current font.
font.set_text(s, 0, flags=flags)
kmeans = KMeans(n_clusters=2)
show_clusters_edge(kmeans,X)
D:\software\anaconda\lib\site-packages\matplotlib\backends\backend_agg.py:240: RuntimeWarning: Glyph 65306 missing from current font.
font.set_text(s, 0.0, flags=flags)
D:\software\anaconda\lib\site-packages\matplotlib\backends\backend_agg.py:203: RuntimeWarning: Glyph 65306 missing from current font.
font.set_text(s, 0, flags=flags)
