Opinion Mining Method for Sentiment Analysis

Abstract

We are living in a world full of data. Every passing second, large data is generated by Social Media, E-Commerce, Stock Exchange and many other platforms. Now-a-days, microblogging sites are used for manypurposes, such as communication, trend detection in marketing and business products, sentiment analysis, election prediction, education and much more, which has changed the public perspective of personalization and socialization. Twitter, is one of the major sources of data which generates more than 500 million of tweets per day. Every user usually shows his feeling or emotion about the topics of his interest. The reason of information gathering is to find out what the people feel. The decurtate length and the highly colloquial nature of tweets render it very difficult to automatically detect the sentiment and thus the requirement of sentiment analysis. Sentiment Analysis and summarization has in recent years caught the attention of many researchers, as on line text analysis is highly beneficial and asked for in various applications. A typical application is product-based sentiment summarization. This multi-document summarizationinforms users about pros and cons of available products. Sentiment analysis allows box office, social media, business analytics, market and FOREX rate prediction, and is also used in recommender system. In the present era sentiment analysis is most interesting research topic in text mining in the field of NLP in which a valuable knowledge extraction from textual data posted on the social media is an onerous task. A new framework has been proposed in this paper to normalize the text and judge the polarity of textual data aspositive, negative or neutral using an ETL (Extract, Transform, and Load) big data tool called Talend. The algorithm developed focusses on parallelism for performance speed and contributes towards the end result by comparing the accuracy with standard data set .