Feature Selection for Cancer Datasets by Modifying Marine Predator Algorithm

This paper's main goal is to modify an efficient heuristic optimization approach for feature selection purpose.Here, Marine Predator Algorithm for Feature Selection (MPAFS) is proposed.With regard to runtime and accuracy, MPAFS has been compared to Particle Swarm Optimization and Differential Evolution.Real datasets for breast, bladder, and colon cancers were gathered from Iraqi hospitals for this study, along with artificial datasets for assessment.For both actual and artificial datasets, we discovered that MPAFS attained the best accuracies with the shortest runtime when compared to other chosen methods.