AIRIME: a scalable adaptive RIME algorithm for feature optimization and data-driven monkeypox disease prediction

The rapid growth of big data in recent applications has created a pressing need for scalable and efficient optimization algorithms capable of handling large-scale and high-dimensional problems. Addressing this challenge, an Adaptive Improved Rime-Ice Optimization algorithm (AIRIME) is proposed which is built on the physics-based RIME algorithm inspired by the formation of soft and hard rime on surfaces. While RIME offers a strong theoretical foundation, it is hindered by limited exploratory capability, slow convergence, and an imbalance between exploration and exploitation. AIRIME introduces four key enhancements, including Adaptive k-Average-Best Mutation (AKAB), Adaptive Searching Process (ASP), Enhanced Solution Quality (ESQ), and Gaussian-Based Optimal Solution Mutation (GBOM), to overcome these challenges and meet the requirements of big data applications. AIRIME was rigorously tested on 41 functions from the CEC2017 and CEC2022 test suites, outperforming various classical, recent and advanced algorithms. Statistical evaluations, including Friedman rank and Wilcoxon rank tests, validated its superior performance and robustness. Sensitivity analyses further demonstrated the adaptability and efficiency of the proposed strategies. Moreover, a binary version of AIRIME was applied to feature selection (FS) tasks across 14 datasets, achieving promising results compared to various established binary metaheuristics. Furthermore, its real-world applicability was demonstrated in a novel Monkeypox disease prediction framework, where AIRIME helped to achieve an exceptional classification accuracy of 95.52% while selecting significantly fewer features, underscoring its efficiency in classifying high-dimensional healthcare data. Thus, AIRIME represents a significant advancement in big data-driven optimization, combining scalability, adaptability, and robustness to address complex and high-dimensional challenges across various domains, including critical healthcare applications.