<![CDATA[Low literacy levels among students pose a significant challenge in supporting academic activities, especially in practical courses in the Information Systems Study Program. This study aims to develop a personalized and relevant book recommendation system using the Neural Collaborative Filtering (NCF) algorithm implemented in TensorFlow and deployed through FastAPI. The dataset used is Book-Crossing, containing over one million user-book interactions. The development follows the CRISP-DM methodology, covering business understanding, data preparation, modeling, and deployment. The NCF model utilizes embedding and dense layers to learn complex user-item interactions. Evaluation shows that the model achieves MAE of 0.3133 and MSE of 0.1531 on training data. The system was successfully deployed and validated through unit testing, capable of providing the top five book recommendations based on user input. The result demonstrates the effectiveness of deep learning approaches in enhancing student literacy through adaptive and integrated recommendation systems.]]>
Copyrights © 2025
