Emotion-Aware Movie Recommendation Systems: Enhancing User Experience through Personalized Entertainment
K Anjana, Sreejisha J Pisharody, Rajalakshmi V R · 2025
With an increasing interest in customized content, systems that recommend movies based on detected emotions are becoming more common. We introduce a novel system known as "Emotion-Aware Movie Recommendation Systems: Personalizing User Experience with Emotional Entertainment," which combines emotion detection with movie recommendation technology.The system offers movie suggestions as per emotions elicited by the users. This is accomplished by incorporating a Bidirectional LSTM (BiLSTM) for emotion extraction and a genre-recommendation variational autoencoder to enhance the quality of user experience. The system extracts emotions from user text inputs, and provides recommendations for movies based on genre scores and IMDb ratings. The BiLSTM model had a validation accuracy of 94.61% and a test accuracy of 94.50%, proving it to be effective in predicting emotions. The recommendation system functioned effectively in providing personalized content, achieving a mean absolute error (MAE) of 0.0189, cosine similarity of 0.9991, and F1-score of 0.9857. This indicates that emotion-aware systems hold promise for delivering more relevant and engaging film recommendations.