OPINION MINING SYSTEM FOR PREDICTING ELECTION OUTCOME USING ASPECT-BASED SENTIMENT ANALYSIS

Abstract

The fascination to predict the future is one of the intent and desires that humanity aim to achieve. These days, general public react to things and share opinions by the means of social media. This behaviour is on daily increase as people tend to divulge feelings and thoughts over the internet but sometimes this information is false and sometimes not useful. This study contributes to the emerging research on sentiment analysis using aspect-based model of social media contents related to a certain political events and trending topics in the political landscape. The system was implemented using Hypertext Preprocessor (PHP), Hypertext Markup Language (HTML), Cascading Style Sheet (CSS). This study focused on two (2) main political parties in Nigeria: All Progressives Congress (APC) and Peoples Democratic Party (PDP). Peoples’ sentimental opinions were gathered (both positive and negative) and analyzed. It tested the difference in the polarity of the sentiments. The result shows that there is a difference in positive and negative sentiments and especially a significant difference in the negative sentiments between the political parties and how it affects the election outcome.