Political Opinion Mining for Popularity Prediction using Machine Learning Techniques
Dharani Devi G, Hemalatha R, R. Pradeep, G C Jagan · 2023
Twitter's utility among voters, politicians, and others keeps rising. This research offers the findings of an investigation into the viability of using Twitter data for purposes such as monitoring public sentiment and anticipating vote preferences. To gauge public sentiment, we amassed and categorised tweets regarding government programmes using natural language processing techniques. We were able to predict the electoral fortunes of various political parties using supervised learning techniques applied to a huge dataset of tweets. Unsupervised learning methods were used to analyse the cleaned data for trends and patterns. To identify whether a tweet was positive or negative towards a candidate, we used supervised learning techniques. Our findings suggest that Twitter data can be utilised to predict the electoral fortunes of political parties and the public's reaction to policy proposals. We ran across problems like the overuse of irony and sarcasm and the usage of difficult-to-decipher jargon and lingo. Because of these obstacles, thorough data pre-treatment and analysis is required before any results can be drawn. Our findings suggest political campaigns and parties might utilise Twitter data to gauge public opinion and attract new voters. Despite some reservations, this essay defends the use of social media data for this.