Local and global approaches of affinity propagation clustering for large scale data - Computer Science > LearningReportar como inadecuado




Local and global approaches of affinity propagation clustering for large scale data - Computer Science > Learning - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Abstract: Recently a new clustering algorithm called -affinity propagation- AP hasbeen proposed, which efficiently clustered sparsely related data by passingmessages between data points. However, we want to cluster large scale datawhere the similarities are not sparse in many cases. This paper presents twovariants of AP for grouping large scale data with a dense similarity matrix.The local approach is partition affinity propagation PAP and the globalmethod is landmark affinity propagation LAP. PAP passes messages in thesubsets of data first and then merges them as the number of initial step ofiterations; it can effectively reduce the number of iterations of clustering.LAP passes messages between the landmark data points first and then clustersnon-landmark data points; it is a large global approximation method to speed upclustering. Experiments are conducted on many datasets, such as random datapoints, manifold subspaces, images of faces and Chinese calligraphy, and theresults demonstrate that the two approaches are feasible and practicable.



Autor: Dingyin Xia, Fei Wu, Xuqing Zhang, Yueting Zhuang

Fuente: https://arxiv.org/







Documentos relacionados