ptimizing Skin Disease Diagnosis using Metaheuristic Algorithms: A Comparative Study

Abstract

Skin diseases, having a wide range of symptoms and appearances, has been putting stern challenge in the field of dermatology. However, early and accurate diagnosis are crucial factors in the field of dermatology to treat and manage skin conditions effectively. In deep demand, the studyreveals the potential of metaheuristic algorithms for skin disease diagnosis and aims a comparison with traditional diagnostic techniques. Areal-time dataset is collected including clinical information, medical images and histopathological data of several patients affected with different skin diseases. The test dataset has beenreviewed to ensure its perfection and representation among several categories of diseases. Several metaheuristic algorithms are introduced like ParticleSwarm Optimization (PSO), Genetic Algorithm (GA), Antlion Optimization (ALO) and Ant Colony Optimization (ACO) in thisstudy. These algorithms are customized for skin disease diagnosis fulfilling all the requirements. To examine the performance of the proposed metaheuristic algorithms, a comparative analysis is conducted. Furthermore, certain performance metrics such as diagnostic accuracy and results of standard deviation, mean fitness score, best fitness score, and worst fitness score are calculated. Theinitial results of this study show that the metaheuristic algorithms have high potentialsfor effective diagnosing of skin diseases. The obtained results are not only delivering highest accuracy but computational speed is also improved. In addition, the conducted comparative analysis also indicatesthe variations in selecting different metaheuristic algorithms. The achieved resultsshowed that the ALO algorithm has outperformed other algorithms with 93% accuracy level. While the ACO achived 90%, the GA has 89% and the PSO worked well with 88% accuracy.