Analysis of data retrieval and opinion mining system

This paper presents analysis of text based data retrieval system and opinion mining on social networking website. This data is collected from various sources like local machine, email accounts, social networking accounts of respective user. Multiple users can use this system by providing log in credentials. This paper explains significance of Inverse Document Frequency and Term Frequency in Lucene scoring formula. Most of the people express their correct reviews on social networking websites than any other discussion forums. This paper discusses an approach to classify each tweet from Twitter into positive, negative or neutral category. This paper also presents techniques to improve performance of Lucene by modifying certain parameters of document scoring formula. Lucene performance also can be improved by modifying algorithm for incremental indexing and parallel processing. The purpose of developing such system is to reduce manual efforts of searching to greater extent. This system is portable and secure.