A Hybrid Multi-Objective Immune Algorithm for Commercial Recommendation Systems

Abstract

Recommender systems are essential components of modern commercial platforms, significantly impacting user engagement and sales. However, balancing accuracy, diversity, and novelty in recommendations remains a challenge. This study presents a novel approach, SVD-MOIA, which combines Singular Value Decomposition (SVD) with the Multi-Objective Immune Algorithm (MOIA) to address this issue. The proposed method aims to enhance recommendation accuracy while simultaneously increasing diversity and novelty. Evaluated on Amazon Product Review, Book-Crossing, and IMDb movies datasets, SVD-MOIA demonstrates superior performance over traditional algorithms. The results show that SVD-MOIA effectively balances multiple conflicting objectives, providing valuable insights for improving user satisfaction and engagement in commercial recommender systems.