SED-Stream: discriminative dimension selection for evolution-based clustering of high dimensional data streams
by Kitsana Waiyamai; Thanapat Kangkachit; Thanawin Rakthanmanon; Rattanapong Chairukwattana
International Journal of Intelligent Systems Technologies and Applications (IJISTA), Vol. 13, No. 3, 2014

Abstract: Clustering of high dimensional data streams become one of the most challenging data mining tasks. Our previous work, SE-Stream is a standard-deviation based projected clustering method to support high dimensional data streams. Besides its ability to find clusters within subgroups of dimensions, SE-Stream is able to monitor and detect change in the clustering structure during the progression of data streams. Extended from SE-Stream, some selected dimensions are used to represent the clusters. Our idea is to select a better set of dimensions to increase the quality of the output clustering. Our proposed SED-Stream projects any cluster to its discriminative dimensions that are highly relevant to the cluster itself but distinguished from the other clusters. Experimental results on both real-world and synthetic stream datasets show that SED-Stream is better than its previous version, SE-Stream, in terms of both purity and f-measure. Compared with HPStream, a state of the art algorithm for projected clustering of high dimensional data streams, SED-Stream outperforms HPStream in terms of f-measure, and has comparable purity.

Online publication date: Wed, 15-Oct-2014

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Intelligent Systems Technologies and Applications (IJISTA):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com