期刊名称：International Journal of Innovative Research in Science, Engineering and Technology
印刷版ISSN：2347-6710
电子版ISSN：2319-8753
出版年度：2015
卷号：4
期号：9
页码：9025
DOI：10.15680/IJIRSET.2015.0409103
出版社：S&S Publications
摘要：With the explosive growth of user generated messages, twitter has become a social site where millionsof users can exchange their opinion. Sentiment analysis on twitter data has provided an economical and effective wayto expose public opinion timely, which is critical for decision making in various domains. For instance, a company canstudy the public sentiment in tweets to obtain users' feedback towards its products; while a politician can adjust his/herposition with respect to the sentiment change of the public. There have been a large number of research studies andindustrial applications in the area of public sentiment tracking and modelling. Millions of users share their opinions onTwitter, making it a valuable platform for tracking and analyzing public sentiment. Such tracking and analysis canprovide critical information for decision making in various domains. Therefore it has attracted attention in bothacademia and industry. Previous researches showed that the tweet was classified appropriately only if the tweet wouldcontain the exact same label (use to detect sentiment) as in the training set. But this approach fails when the tweetcontains a synonym or a variant of the label (having same meaning) instead of the exact same label. Although thetweet should have been classified accurately because the variant in the tweet and the label in the training set had samemeaning. To solve this problem, a Lexicon based approach using naive bayes classifier for automatic analysis oftwitter message is presented. The results show that by incorporating a lexicon based approach with the bayes classifier,the efficiency and the accuracy of the classifier to classify the tweets has improved significantly.
