Landmark selection for spectral clustering based on Weighted PageRank

详细信息    查看全文

• 作者：D. Rafailidis ; ^{draf@csd.auth.gr}Author Vitae ; E. Constantinou ^{econst@csd.auth.gr}Author Vitae ; Y. Manolopoulos ^{manolopo@csd.auth.gr}Author Vitae
• 关键词：Spectral clustering ; Sparse coding ; Databases
• 刊名：Future Generation Computer Systems
• 出版年：2017
• 出版时间：March 2017
• 年：2017
• 卷：68
• 期：Complete
• 页码：465-472
• 全文大小：792 K
• 卷排序：68

文摘

We select representative landmarks with weighted PageRank for spectral clustering. Scalability in spectral clustering is ensured by following a landmark strategy. We examine several landmark selection strategies for spectral clustering. We experimentally show the superiority of the proposed method over competitors. We provide future directions in spectral clustering.