An improved approach for online trending forum detection based on sentiment analysis

Sentiment analysis is all about opinion mining. Analyze emotions, characteristics and opinions of users about any products, topics, or problem. For the public opinion, internet is becoming a spreading and very wide platform where online forums, social sites, blogs and other sites contains opinion and review of people in form of comments and posted messages. Now a days the knowledge obtained from these sites, blogs and comments and poster is very useful for marketing research. In these process to detect online trending forum, In first phase emotional polarity of poster messages is calculated and in second phase trending forums are identified using clustering method i.e. EM (Expectation Maximization) and different classification methods i.e. SVM, PSO-SVM, Naïve Bayes, Random forest. In this research paper, emotional polarity of poster comments calculated by Lexicon approach of sentiment analysis using python script. After that trending forums are identified by two approaches EM clustering and Random Forest classification technique gives better result than previous approaches to detect trending forum.