QQ客服 官方微博 反馈留言

	A new markov model for clustering categorical sequences
	Xiong tengke; Wang shengrui; Jiang qingshan; Huang Joshazhexue
	2011
会议名称	11th IEEE International Conference on Data Mining
会议地点	Sydney
英文摘要	Clustering categorical sequences remains an open and challenging task due to the lack of an inherently meaningful measure of pairwise similarity betweensequences. Model initialization is an unsolved problem in model-based clustering algorithms for categorical sequences. In this paper, we propose a simple and effective Markov model to approximate the conditional probability distribution (CPD) model, and use it to design a novel two-tier Markov model to represent asequence cluster. Furthermore, we design a novel divisive hierarchical algorithm for clustering categorical sequences based on the two-tier Markov model. The experimental results on the data sets from three different domains demonstrate the promising performance of our models and clustering algorithm. © 2011 IEEE.(22 refs)
收录类别	EI
语种	英语
内容类型	会议论文
源URL	[http://ir.siat.ac.cn:8080/handle/172644/3582]
专题	深圳先进技术研究院_数字所
作者单位	2011
推荐引用方式 GB/T 7714	Xiong tengke,Wang shengrui,Jiang qingshan,et al. A new markov model for clustering categorical sequences[C]. 见:11th IEEE International Conference on Data Mining.  Sydney.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。