IJAIEM

Malluri Mounika,Pariveda Cherishma , Vankayalapati Rikitha ,T Suresh

Abstract

Public opinion polling plays a crucial role in political analysis and electoral forecasting. Traditional polling methods are expensive, time-consuming, and often biased due to improper population sampling. This paper proposes a novel approach using tweet sentiment analysis for opinion polling. The proposed system leverages natural language processing (NLP) and machine learning algorithms to analyze social media data, specifically political tweets, to estimate public sentiment and predict electoral outcomes. The system incorporates sentiment classification, heuristic estimation, and hybrid regression models to enhance accuracy and reliability. Experimental results show that this approach provides more frequent and cost-effective public opinion insights with higher accuracy compared to conventional polling methods.