Rival Penalized Competitive Learning for Model-Based Sequence Clustering (2000)

Content Provider	CiteSeerX
Author	Law, Martin H. Kwok, James T.
Description	In this paper, we propose a model-based, competitive learning procedure for the clustering of variable-length sequences. Hidden Markov models (HMMs) are used as representations for the cluster centers, and rival penalized competitive learning (RPCL), originally developed for domains with static, fixed-dimensional features, is extended. State merging operations are also incorporated to favor the discovery of smaller HMMs. Simulation results show that our extended version of RPCL can produce a more accurate cluster structure than k-means clustering. 1. Introduction Clustering aims at dividing a set of observations into different groups so that members of the same group are more alike than members of different groups. Knowledge of this cluster structure is usually very important in understanding data with unknown distributions. Traditional clustering algorithms work only with data having static, fixed-dimensional features. However, with the increasingly widespread use of information sys...
File Format	PDF
Language	English
Publisher Date	2000-01-01
Publisher Institution	In Proceedings of the Fifteenth International Conference on Pattern Recognition
Access Restriction	Open
Subject Keyword	Different Group Variable-length Sequence Model-based Sequence Clustering Rival Penalized Competitive Learning Cluster Center Widespread Use Competitive Learning Procedure Simulation Result Fixed-dimensional Feature Hidden Markov Model Information Sys Unknown Distribution State Merging Operation Accurate Cluster Structure Algorithm Work Cluster Structure Extended Version
Content Type	Text
Resource Type	Article