Sentiment Analysis on Effective Governance using Deep Learning

Abstract

Public opinion plays a crucial role in evaluating the performance and effectiveness of governance. With the rapid growth of social media, citizens freely express satisfaction, criticism, and expectations regarding government policies. This research proposes a deep learning-based sentiment analysis framework to quantify and interpret public perception of governance. Using a curated dataset from social platforms, news portals, and public forums, the study implements LSTM, Bi-LSTM, and BERT to classify sentiments into positive, negative, and neutral categories. Experimental results show that transformer-based models outperform recurrent neural architectures, achieving an accuracy of 92.6%. The findings highlight how deep learning can provide actionable insights for policy formulation, crisis response, and public communication strategies.

Keywords