Web-Page Recommendation using Fuzzy Ontology with Public Sentiment Variations on Twitter
Twitter is a social site where millions of users can exchange their opinion with the explosive growth of user generated messages. Decision making becomes critical in various domains, so sentiment analysis of Twitter data provides an economical and effective way to the people who can expose opinion timely. Because of the opinions shared by Millions of users, twitter becomes a valuable platform for tracking and analyzing public sentiment. Therefore attention of people increases in both academia and industry. Based on this observation, Latent Dirichlet Allocation (LDA) based model is proposed, in which Foreground and Background LDA (FB-LDA) is used. Foreground topics are distilled and filtered out longstanding background topics. Readability of the mined reasons is enhanced to develop another generative model called Reason Candidate and Background LDA (RCB-LDA) in which ranking procedure is implemented with respect to their “popularity” within the variation period. The proposed work considers the fuzzy ontology, a novel method is proposed to efficiently provide better Web-page recommendation by using semantic-enhancement which is integrating the domain and Web usage knowledge of a website. The domain knowledge is represented by two new models. The first one uses ontology to represent the domain knowledge. The second model uses the semantic network which is generated automatically to represent domain terms, Web-pages, and the relations between them.