一般相似度函數下譜聚類的收斂性

Translated title of the contribution: Convergence of spectral clustering with a general similarity function

高煒, 周定軒*

*Corresponding author for this work

Research output: Journal Publications and Reviews › RGC 21 - Publication in refereed journal › peer-review

Abstract

譜聚類算法由與相似度函數相關的圖Laplace算子的特徵函數產生.本文證明與一般相似度函數相關的譜聚類算法的收斂性,並使用覆蓋數方法對收斂性給出量化估計.當相似度函數是歐氏空間子集上一個Lipschitzs>0函數時,O((log(n+1))～(1/2)/n～(1/2))形式的收斂率得到證實.我們同時指出一個相應函數集的覆蓋數的增長可以表現任意差.
Translated title of the contributionConvergence of spectral clustering with a general similarity function
Original languageChinese (Traditional)
Pages (from-to)985 - 994
JournalScience China Mathematics
Volume42
Issue number10
DOIs
Publication statusPublished - 2012

Access Output

Other Access Options

Check CityUHK Library

Permanent Link

Cite this