Opinion Mining of Live Comments from Website using Fuzzy Logic and NLP

For many Natural Language Processing tasks, Opinion Mining of text content is important. In recent years, Social Media plays important role for expressing and sharing of any valuable or important information in terms of text, SMSs, mails, reviews or comments, etc. Existing studies of Opinion Mining tend to extract less features and also used static datasets i.e. publicly available datasets. So the proposed system gives a heuristic approach for Dynamic Comment Classification. The approach focuses on crawling of web pages by entering seed URL and then parsing of web pages is done. After getting user and their comments, preprocessing of comments is done. Various features are extracted like term weight, Noun Identification, Thematic words, Bag of words, etc. Then by applying Fuzzy Logic and IF-THEN rules, comment classification is done: positive and negative. Evaluation is done by the evaluation parameters like Precision, recall and F-measure.