A Proposed Model for Enhancing Product Recommendation Based on Word Embedding
Omar Attia, Ahmed Dahroug, Abdel-Fattah Hegazy · 2023
Online commerce is a thriving sector, marked by millions of daily transactions across the globe. This paper introduces a model for product recommendation grounded in customer behavior within e-commerce platforms. Leveraging data from customer profiles, the model incorporates purchase history, product ratings, and browsing patterns to construct comprehensive user and item profiles. Utilizing word embedding techniques, latent factors and patterns in user-item interactions are identified. To address the complexity of user behavior, model facilitates both random forest and tree ensemble models are integrated. The proposed highly personalized and relevant product recommendations, enhancing user satisfaction and driving sales. Experimental validation demonstrates the model’s effectiveness compared to existing approaches, emphasizing the role of product annotation enhancements.