Elucidating the Public Sentiment Fluctuations on Twitter

Now a days numbers of people share their opinion on different aspect of life every day. So, twitter is used as microblogging platform for communicate with each other and also tracking and evaluating public emotions. It facilitates the details for making decisions in different domain. It is an attraction for organization and in academics field. Antecedently, research concentrated modelling and tracking public emotions. Here we proceed for the next step that is elucidating sentiment fluctuations. We have to take recent topic within the sentiment fluctuation periods which are strongly linked to the reason behind fluctuations. Grounded on this, we propose a Latent Dirichlet Allocation (LDA) base model. Foreground and Background LDA (FB-LDA), take the recent (foreground) topics and separate out long-lasting (background) topics. These recent topics give the possible elucidation of sentiment fluctuations. For better readability, we choose the illustrative tweets for recent topics and trained another productive model addressed Reason Candidate and Background LDA (RCB LDA). For ranking this model according to their popularity within the fluctuation period. This method can detect recent topics and rank reason candidates.