Please use this identifier to cite or link to this item:
Title: An Efficient Feature Selection Technique of Unsupervised Learning Approach for Analyzing Web Opinions
Authors: Valli, M S
Arasu, G T
Keywords: Feature Based Summarization;Feature Selection;Support Vector Machine;Rough Set Theory
Issue Date: Apr-2016
Publisher: NISCAIR-CSIR, India
Abstract: Examination of developing Web opinions is probably valuable for realize enduring topics of public like crime and terrorist attack detection, Some participants or users are using the web forum to publicize their thought about particular incident in the world such as committing crime. How the topics are progressed together with the social interaction between contributors and recognize important and influence of participants discussions of various topics through social media. Participants are usually considering opinions to be articulated through adjectives, and make wide use of moreover general dictionaries or specialist to provide the appropriate adjectives.  However, analyzing and clustering of Web opinions is really difficult. Unlike the documents of Web opinions are tiny and sparse with noisy text content, typical Web opinions document clustering method produce unsatisfactory performance. Feature selection (FS) is a procedure which efforts to select more informative features. Some Web opinions documents have too many repeated and irrelevant texts for classification or clustering. Feature selection method can recover this problem instead of classification and clustering algorithms. The main aspect of feature selection is give high accuracy performance with minimal feature subset. In this paper, we propose the unsupervised rough set method for clustering text for Web opinion mining. We conducted more experiments and had benchmarked with the unsupervised algorithm which gives higher micro accuracy results.
Page(s): 221-224
ISSN: 0975-1084 (Online); 0022-4456 (Print)
Appears in Collections:JSIR Vol.75(04) [April 2016]

Files in This Item:
File Description SizeFormat 
JSIR 75(4) 221-224.pdf226.02 kBAdobe PDFView/Open

Items in NOPR are protected by copyright, with all rights reserved, unless otherwise indicated.