谱聚类（spectral clustering）

最新推荐文章于 2024-01-11 22:08:31 发布

原创最新推荐文章于 2024-01-11 22:08:31 发布 · 288 阅读

0 ·

CC 4.0 BY-SA版权

数据挖掘专栏收录该内容

5 篇文章

订阅专栏

本文介绍了一个使用谱聚类算法进行数据聚类的实际案例。通过读取特定格式的数据文件，对数据进行预处理，并利用高斯核函数计算相似度矩阵，最终实现基于谱聚类的数据划分并可视化展示。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

谱聚类（spectral clustering）点击打开链接

import matplotlib.pyplot as plt
import numpy as np
from numpy import linalg as LA
from sklearn.cluster import KMeans
from sklearn.datasets import make_blobs
from sklearn.metrics.pairwise import rbf_kernel
from sklearn.preprocessing import normalize
import tensorflow as tf
from mpl_toolkits.mplot3d import Axes3D


data = []

cty_data = []
def data_p(fliename):
    fr = open(fliename)

    tid = 0
    cid = 0
    for line in fr.readlines():     
        tline = line.strip().split(';')

        data.append(tline)
        
        cn = tline[0]
        if tid==0:
            cty_data.append([tid])
            #print(cty_data)
            #print(cty_data[cid][0])
        elif cn == data[cty_data[cid][0]][0]:
            cty_data[cid].append(tid)
        else:
            cid += 1
            cty_data.append([tid])

        tid += 1
            

data_p('C:\\Users\\imac\\Desktop\\2018\\bigdata\\py\\wealth1951.txt')
cty_data = np.asarray(cty_data)
#print(cty_data)
data = np.asarray(data)
#print(data[cty_data[:,:],0])

cdata = []

def getdata():
    for i in range(len(cty_data)): #国家
        itm =  data[cty_data[i,:],3:6]
        #print(itm)
        itm = np.asarray(itm)
        itmm = []
        for j in range(3): #维数
            for k in range(len(itm)-1,0,-1):  #年份从后往前
                itm[k,j] = (float(itm[k,j])-float(itm[k-1,j]))/float(itm[k-1,j])
                #变化率
            #print(itm[1:len(itm),j])
            itmm.append(np.mean(np.asarray(itm[1:len(itm),j],dtype=np.float32)))
            #每个国家变化率均值
        #print(itm)
        cdata.append(itmm)

   
getdata()
cdata = np.asarray(cdata,dtype=np.float32)


cdata = normalize(cdata)
print(cdata)
def similarity_function(points):

    res = rbf_kernel(points)
    #print(res)
    for i in range(len(res)):
        res[i,i] = 0
    return res

def spectral_clustering(points,k):
    W = similarity_function(points)
    D = np.diag(np.sum(W,axis=1))
    #print(W)
    Dn = np.sqrt(LA.inv(D))
    #L=Dn*(D-W)Dn= I-Dn*W*Dn
    L = np.eye(len(points)) - np.dot(Dn, np.dot(W,Dn))
    eigvals, eigvecs = LA.eig(L)
    indices = np.argsort(eigvals)[:k]  
    k_smallest_eigenvectors = normalize(eigvecs[:,indices])

    return KMeans(n_clusters=k).fit_predict(k_smallest_eigenvectors)


labels = spectral_clustering(cdata, 5)

fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')  
ax.scatter(cdata[:, 0], cdata[:, 1],cdata[:,2], c=labels) 

plt.show()

print(labels)