多元统计分析——层次聚类方法_层次聚类代码案例

from scipy.cluster.hierarchy import dendrogram, linkage,fcluster
import pandas as pd
import numpy as np
from matplotlib import pyplot as plt
from pylab import mpl
mpl.rcParams['font.sans-serif'] = ['SimHei']
from sklearn.preprocessing import MinMaxScaler
from sklearn.preprocessing import StandardScaler
#
data = pd.read_csv(r"WBClust2013.csv",index_col = 'Country')
print(data.head(5))
 
 #极差标准化变换
mm = MinMaxScaler()
mm_data = mm.fit_transform(data)#归一化后的数据
 
 #标准化变换
ss = StandardScaler()
std_data = ss.fit_transform(data)#归一化后的数据
# origin_data = ss.inverse_transform(std_data)#原始数据
 
 #层次信息
z0=linkage(mm_data, method='average', metric='euclidean') #平均距离
z1=linkage(std_data, method='average', metric='euclidean') #平均距离
 
 #类个数
f=fcluster(z0, t=4, criterion='maxclust')#'inconsistent’, ‘distance’
f=fcluster(z1, t=4, criterion='maxclust')#'inconsistent’, ‘distance’
 
fig = plt.figure(figsize=(5, 3))
plt.title('极差标准化')
dn = dendrogram(z0)
plt.show()
plt.title('标准化')
dn = dendrogram(z1)
plt.show()
 
 
 
#第二题
data = pd.read_csv(r"NASAUnderstory.csv",index_col = 'Overstory Species')
print(data.head(5))
#极差标准化变换
mm = MinMaxScaler()
mm_data = mm.fit_transform(data)#归一化后的数据
 
#标准化变换
ss = StandardScaler()
std_data = ss.fit_transform(data)#归一化后的数据
 
#层次信息
z0=linkage(mm_data, method='average', metric='euclidean') #平均距离
z1=linkage(std_data, method='average', metric='euclidean') #平均距离
 
#类个数
f=fcluster(z0, t=4, criterion='maxclust')
f=fcluster(z1, t=4, criterion='maxclust')
 
fig = plt.figure(figsize=(5, 3))
plt.title('极差标准化')
dn = dendrogram(z0)
plt.show()
plt.title('标准化')
dn = dendrogram(z1)
plt.show()

方法二

from scipy.cluster.hierarchy import dendrogram, linkage,fcluster
from matplotlib import pyplot as plt
import pandas as pd
import numpy as np
from sklearn.preprocessing import StandardScaler
 
data=pd.read_csv(r"./WBClust2013.csv",header=0,index_col=0)
print('--------------作业1--------------')
# 数据预处理
ss = StandardScaler()
data = ss.fit_transform(data)#归一化后的数据
 
# 模型训练
metrics=['Euclidean']
# 'mantattan','cosine','chebyshev'
# 选择不用的距离度量:
# 欧氏距离 曼哈顿距离 余弦相似性 切比雪夫距离
Z = linkage(np.array(data), method='ward',metric='euclidean')
f = fcluster(Z,4,'distance')
fig = plt.figure(figsize=(14, 8))
dn = dendrogram(Z)
plt.title(f"method=ward,metric=euclidean")
plt.show()
    
# 选择不同的类间距离,距离度量标准为欧式距离
methods=['single','complete','average','ward']
for th in methods:
    Z = linkage(np.array(data), method=th,metric='euclidean')
    f = fcluster(Z,4,'distance')
    fig = plt.figure(figsize=(14, 8))
    dn = dendrogram(Z)
    plt.title(f"method={th},metric=euclidean")
    plt.show()
 
 
print('--------------作业2--------------')
data1=pd.read_csv(r"./NASAUnderstory.csv",header=0,index_col=1)
 
data1=data1.drop('Labels', axis=1)
data1 = ss.fit_transform(data1)
methods=['single','complete','average','ward']
for th in methods:
    Z = linkage(np.array(data1), method=th,metric='euclidean')
    f = fcluster(Z,4,'distance')
    fig = plt.figure(figsize=(14, 8))
    dn = dendrogram(Z)
    plt.title(f"method={th},metric=euclidean")
    plt.show()
 
'''
from scipy.cluster.hierarchy import dendrogram, linkage,fcluster
from matplotlib import pyplot as plt
X = [[1,2],[3,2],[4,4],[1,2],[1,3]]
Z = linkage(X, 'ward')
f = fcluster(Z,4,'distance')
fig = plt.figure(figsize=(5, 3))
dn = dendrogram(Z)
plt.show()
'''