Binary Starling Murmuration Optimizer Algorithm to Select Effective Features from Medical Data

Feature selection is an NP-hard problem to remove irrelevant and redundant features with no predictive information increase the performance of machine learning algorithms. Many wrapper-based methods using metaheuristic algorithms have been proposed select effective features. However, they achieve differently on medical data, most them cannot find those that may fulfill required accuracy in diagnosing important diseases such as Diabetes, Heart problems, Hepatitis, Coronavirus, which are targeted datasets this study. To tackle drawback, algorithm needed can strike a balance between local global search strategies selecting from datasets. In paper, new binary optimizer named BSMO proposed. It based newly starling murmuration (SMO) has high ability solve different complex engineering it expected also effectively optimal subset Two distinct approaches utilized by when searching Each dimension continuous solution generated SMO simply mapped 0 or 1 variable threshold second approach, whereas first, versions developed several S-shaped V-shaped transfer functions. The was evaluated four datasets, results were compared well-known terms metrics, including fitness, accuracy, sensitivity, specificity, precision, error. Finally, superiority statistically analyzed Friedman non-parametric test. statistical experimental tests proved attains better comparison competitive ACO, BBA, bGWO, BWOA for