Modelling Personalized Book Recommendation using Hybrid Filtering

The paper introduces a new method for personalised book recommendations using a hybrid filtering approach. Traditional methods like Collaborative filtering and Content-based filtering have limitations in capturing user preferences fully. To address this, the study suggests combining both methods to improve accuracy and coverage. The research involves collecting book ratings and metadata, analyzing user preferences, and building a neural network model using TensorFlow. Results show that the hybrid filtering model outperforms individual methods, improving precision, recall, and F1 score. It also handles diverse user preferences and the cold-start problem effectively. Overall, this research enhances recommendation systems by proposing a tailored approach for book recommendations. It provides useful insights for practitioners in developing more user-centric systems. Future research could explore additional data sources, incorporate contextual information, and improve model interpretability for enhanced accuracy and user satisfaction.