A bio-medical snake optimizer system driven by logarithmic surviving global search for optimizing feature selection and its application for disorder recognition

• DOI: 10.1093/jcde/qwad101
• Publication Date: 2023-11-11T08:49:04Z
• Authors: Ruba Abu Khurma; Esraa Alhenawi; Malik Braik; Fatma A Hashim; Amit Chhabra; Pedro A Castillo
• Keywords: Overfitting; Feature (linguistics)

Abstract It is of paramount importance to enhance medical practices, given how important it protect human life. Medical therapy can be accelerated by automating patient prediction using machine learning techniques. To double the efficiency classifiers, several preprocessing strategies must adopted for their crucial duty in this field. Feature Selection (FS) one tool that has been used frequently modify data and classification outcomes lowering dimensionality datasets. Excluded features are those have a poor correlation coefficient with label class, i.e., they no meaningful do not indicate where instance belongs. Along recurring features, which show strong association remainder features. Contrarily, model being produced during training harmed, classifier misled presence. This causes overfitting increases algorithm complexity processing time. The pattern made clearer FS, also creates broader lower chance an acceptable amount time algorithmic complexity. optimize FS process, building wrappers employ metaheuristic algorithms as search algorithms. best solution, reflects subset within particular dataset aids diagnosis, sought study Snake Optimizer (SO). swarm-based approaches SO founded on left general flaws, like local minimum trapping, early convergence, uneven exploration exploitation, convergence. By employing cosine function calculate separation between present solution ideal logarithm operator was paired better exploitation process get over these restrictions. In order overall answer, forces solutions spiral downward. Additionally, employed put evolutionary algorithms’ preservation premise into practice. accomplished utilizing three alternative selection systems – tournament, proportional, linear improve phase. These allow found more thoroughly relation chosen than at random. Tournament Logarithmic (TLSO), Proportional Optimizer, Linear Order Optimizer. A number 22 reference datasets were experiments. findings that, among 86% datasets, TLSO attained accuracy, 82% feature reduction. terms standard deviation, noteworthy reliability stability. On basis running duration, is, nonetheless, quite effective.