Enhancing Machine Learning for Employee Satisfaction Prediction Using MA-SOM: A Mayfly Algorithm Optimized Approach

This paper introduces a novel approach, MA-SOM, which optimizes Self-Organizing Maps (SOM) using the Mayfly Algorithm (MA). Traditional SOMs often encounter challenges related to convergence accuracy and speed, motivating the development of enhanced variants. MA-SOM addresses these issues by leveraging MA's global and local search capabilities to optimize key network parameters such as competition layer dimensions, step size during classification, and tuning of neighborhood distances. Comparative tests against Whale Optimization Algorithm (WOA) and Particle Swarm Optimization (PSO) demonstrate MA's effectiveness in achieving improved convergence speed and precision. Practical application of MA-SOM is showcased through a case study on employee satisfaction classification in the tourism industry, highlighting its superior performance in accuracy, fitness values, and convergence speed. This work presents MA-SOM as a promising solution for optimizing SOMs and addressing real-world data analysis challenges.