A Similarity Measure for Text Classification and Clustering |
Author(s): |
P. Mayilsamy , GOBI ARTS & SCIENCE COLLEGE; P. Elango, GOBI ARTS & SCIENCE COLLEGE |
Keywords: |
Clustering Algorithms, Clustering Applications, Similarity Measures, Text Clustering |
Abstract |
This paper introduces a live of similarity between two clustering’s of constant dataset made by two completely different algorithms, or maybe constant algorithmic rule. Mensuration the similarity between documents is a crucial operation within the text process field. This paper projected a replacement similarity measure. To compute the similarity between two documents with relevance a feature, the proposed measure takes the subsequent three cases into account: a) The feature seems in each documents, b) The feature seems in just one document, and c) The feature seems in none of the documents. For the primary case, the similarity will increase because the distinction between the two concerned feature values decreases. Moreover, the contribution of the distinction is generally scaled. For the second case, a hard and fast worth is contributed to the similarity. For the last case, the feature has no contribution to the similarity. The effectiveness of the live is evaluated on many real-world knowledge sets for text classification and bunch issues. |
Other Details |
Paper ID: IJSRDV3I80151 Published in: Volume : 3, Issue : 8 Publication Date: 01/11/2015 Page(s): 563-565 |
Article Preview |
|
|