Deep Learning Approach to Tweet Sentiment Analysis for Movie Recommendation Systems

Abstract

A deep learning approach to analyze the sentiment of user tweets is proposed to provide recommendations for movies and TV shows on streaming services. Recognizing the importance of implicit user feedback on social media, we focus on how viewers express their opinions on Twitter regarding drama series and films. To achieve this, we collect Twitter data using a Python library and REST API, then employ natural language processing techniques like TextBlob and VADER to extract show names and analyze sentiment. Our deep learning model demonstrates positive results with strong accuracy and minimal error rates. The proposed approach targets social media users who also subscribe to streaming services and opens potential applications in diverse domains like purchasing preferences, political opinions, and educational choices.