DMCA
A NOVEL APPROACH FOR HIGH DIMENSIONAL DATA CLUSTERING
| Citations: | 1 - 0 self |
BibTeX
@MISC{Tidke_anovel,
author = {B. A Tidke and R. G Mehta and D. P Rana},
title = {A NOVEL APPROACH FOR HIGH DIMENSIONAL DATA CLUSTERING},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
High dimensional data clustering is the analysis of data with few to hundreds of dimensions. Large dimensions are not easy to handle and impossible in certain cases to visualize. To improve the efficiency and accuracy of clustering on high dimensions, data reduction is required as pre-processing. A clustering ensemble is a paradigm that combines the outputs of several clustering algorithms to achieve a more accurate and stable final output. Clustering ensemble method based on a novel two-staged clustering algorithm is proposed in this paper. Existing clustering techniques normally merge small cluster with big ones results in removing the identity of those small clusters. The proposed algorithms work on split and merge technique to overcome this limitation. Experimental results of the proposed method on several data sets are compared with individual clustering results produced by well-known clustering algorithms.
Keyphrases
small cluster novel two-staged clustering algorithm stable final output large dimension high dimensional data clustering algorithm work big one result several data set data reduction merge technique well-known clustering algorithm high dimension ensemble method individual clustering result experimental result clustering ensemble certain case
