Multi-Objective Feature Selection: Reviewing Decades of Research

This comprehensive review explores MultiObjective Feature Selection (MOFS), a crucial technique for optimizing performance and classification accuracy in highdimensional datasets. MOFS addresses the challenge of selecting relevant features while balancing conflicting objectives, such as minimizing feature count and maximizing model performance. The study explores various MOFS approaches, including filter, wrapper, embedded, and hybrid methods, each with distinct advantages and limitations. Key findings highlight the diversity of MOFS techniques and their effectiveness in handling complex, real-world applications like cancer detection. The review also identifies ongoing challenges in the field, such as scalability issues, computational costs, and the need for improved evaluation measures. By providing an in-depth analysis of current MOFS methods and their applications, this article serves as a valuable resource for researchers and practitioners, paving the way for future advancements in Feature Selection across various domains.