Concept Detection and Cluster Analysis from Newsfeed-Singular Value Decomposition Based Approach

Abstract

Concept detection plays an important role if there is a huge amount of data available. We know that cluster analysis, topic detection, opinion mining have got a major role in the product marketing, online shopping, E-commerce. In this paper, we have conducted the topic detection and clustering experiments on the News samples which were sourced from online newspapers. Our aim is to find out the topics which also available in the text documents as a group of words and apply a clustering technique using the Singular value decomposition method. Then opinions are extracted from the comments, collected on a particular subject of interest like the comments for Smartphone. Finally, the clustering technique is applied on these sentiments to figure out the opinions of the people towards different features of the Smartphone. The results obtained here are competitive with the technology available.