Temporal Sentiment Analysis and Causal Rules Extraction from Tweets for Event Prediction

Abstract Sentiment analysis or opinion mining is the process of computationally identifying and categorizing opinions expressed in a piece of text, in order to determine whether the writer's attitude towards a particular topic, product, etc. is positive, negative, or neutral. It is one of the most active research areas in natural language processing and text mining in recent years. A detailed study of the two concepts (1) Temporal sentiment analysis (2) Sentiment causal relation is presented in this paper. Temporal sentiment analysis is useful for summarizing the events based on sentiment and time. Causal relation is useful for identifying cause and effect of events and is also useful for event prediction. These two concepts when combined result in a better event prediction model that can predict the time period between the events and sentiment of upcoming events. The proposed work introduces a generalized prediction model based on temporal sentiment analysis of tweet to identify the causal relation between the events which can be used to predict the event sentiment and duration between the events. The proposed method is to be evaluated using the performance measures precision and recall. The accuracy of causal rule prediction is evaluated using parameters Mean Absolute Error (MAE) and the Root Mean Squared error (RMSE).