统计学习理论及应用 第十讲非监督学习 编写：文泉、陈娟 电子科技大学 计算机科学与工程学院
统计学习理论及应用 第十讲 非监督学习 编写：文泉、陈娟 电子科技大学 计算机科学与工程学院
主目录 ①K-Means聚类算法(Clustering Algorithm) 2 网页排序PageRank 1/32
主目录 1 K-Means 聚类算法 (Clustering Algorithm) 2 网页排序 PageRank 1 / 32
一级目录 ①K-Means聚类算法(Clustering Algorithm) ②网页排序PageRank 2/32
一级目录 1 K-Means 聚类算法 (Clustering Algorithm) 2 网页排序 PageRank 2 / 32
l.K-Means聚类算法(Clustering Algorithm) In the clustering problem,we are given a training set {x(),...x(,x()ER", and want to group the data into a few cohesive"clusters."No labels(are given. The k-means clustering algorithm is as follows: Initialize cluster centroidsE R",j=1,2,...,k,randomly. 2 Repeat until convergence: For each training samplex(,set c0=argmin|lr0-马l2 2 For every cluster centroid i,set ∑l(c0=r0 ∑1I(c0=) } I()is the indicator function. 3/32
1. K-Means 聚类算法 (Clustering Algorithm) ▶ In the clustering problem, we are given a training set {x (1) , . . . , x (m)}, x (i) ∈ R n , and want to group the data into a few cohesive “clusters.” No labels y (i) are given. ▶ The k-means clustering algorithm is as follows: 1 Initialize cluster centroids µj ∈ R n , j = 1, 2, . . . , k, randomly. 2 Repeat until convergence: { 1 For each training sample x (i) , set c (i) = arg min j ∥x (i) − µj∥2 2 For every cluster centroid µj , set µj = Pm i=1 I(c (i) = j)x (i) Pm i=1 I(c (i) = j) } • I(·) is the indicator function. 3 / 32
一、二级目录 ① K-Means聚类算法(Clustering Algorithm) 。解释Explanation o收敛性Convergence of K-Means o矩阵建模Matrix Modelling of K-Means 4/32
一、二级目录 1 K-Means 聚类算法 (Clustering Algorithm) 解释 Explanation 收敛性 Convergence of K-Means 矩阵建模 Matrix Modelling of K-Means 4 / 32