Efficient evolutionary spectral clustering

详细信息    查看全文

• 作者：Rocco Langone ; ^a ; ^{rocco.langone@esat.kuleuven.be} ; Marc Van Barel^b ; Johan A.K. Suykens^a
• 关键词：Evolutionary spectral clustering ; Incomplete Cholesky decomposition ; Linear complexity ; Temporal smoothness
• 刊名：Pattern Recognition Letters
• 出版年：2016
• 出版时间：1 December 2016
• 年：2016
• 卷：84
• 期：Complete
• 页码：78-84
• 全文大小：1461 K
• 卷排序：84

文摘

An evolutionary spectral clustering method using a smoothed Laplacian is presented. The incomplete Cholesky decomposition (ICD) reduces runtime from O(N3) to O(N). The memory requirements are decreased from O(N2) to O(N). A stopping criterion using the convergence of cluster assignments is adopted. An efficient procedure to select the number of clusters is presented.