Personalized Book Recommendation Using Artificial Neural Networks Based on User Behavior Analysis

Abstract — Books are available extensively in libraries, which creates a problem in finding a suitable book to read. Therefore, personalized book recommendation systems have been designed using machine learning techniques to recommend a particular book to the user. The project presents a Book Recommendation System that uses an Artificial Neural Network (ANN) for intelligent, personalized book recommendation. proposed The system takes in book data such as title, author, genre, and ratings, and runs an ANN model created with TensorFlow that generates recommendations based on those. A new system for personalized recommendations based on user interest is content-based filtering with neural network predictions for neural network predictions. Implementation of the system consists of preprocessing methods, such as one-hot encoding for category variables, imputation of missing values, and feature transformation for optimizing performance of the model. The system is developed in Django to furnish a web interface user-friendly and accessible through easy-endpoint APIs. Experimental studies have confirmed improvement in the recommendation accuracy and user satisfaction, thus further testifying to the supremacy of artificial neural networks (ANNs) in improving the process of book discovery.