A novel E-commerce recommender system using deep learning approaches*
Ikram Karabila, Nossayba Darraz, Anas El-Ansari, Nabil Alami, Salma Chenouni, Mostafa El Mallahi · 2024
Recommender systems are essential for optimizing user experience and engagement within e-commerce platforms. This study proposes a novel deep learning-based approach for recommendation systems, integrating GloVe, Convolutional Neural Networks (CNN), Word2Vec, Multi-Layer Perceptron (MLP), and SVD++. With a focus on addressing prevalent challenges, particularly data sparsity, this system aims to improve the accuracy and personalization of e-commerce recommendations. Through comprehensive evaluation on a real-world dataset, our approach demonstrates its effectiveness in delivering personalized recommendations on e-commerce platforms, achieving an impressive RMSE of 0.509, and MAE of 0.595. By combining deep learning, collaborative filtering, and ensemble techniques, our system substantially enhances performance, leveraging additional insights extracted from product metadata and user reviews.