Improvement of grey wolf optimizer with adaptive middle filter to adjust support vector machine parameters to predict diabetes complications

Abstract In medical science, collecting and classifying data from various diseases is a vital task. The confused large amounts of are problems that prevent us achieving acceptable results. One the major for diabetic patients failure to properly diagnose disease. As result this mistake in diagnosis or early diagnosis, patient may suffer complications such as blindness, kidney failure, cutting off toes. Nowadays, doctors disease by relying on their experience knowledge performing complex time-consuming tests. with current diabetic, diagnostic methods lack appropriate features consequently weakness its especially stages. Since diabetes relies many parameters, it necessary use machine learning support vector (SVM) predict diabetes. disadvantages SVM parameter adjustment, which can be accomplished using metaheuristic algorithms particle swarm optimization algorithm (PSO), genetic algorithm, grey wolf optimizer (GWO). paper, after preprocessing preparing dataset mining, we based selected parameters acquired laboratory test improved GWO. We improve selection process GWO employing dynamic adaptive middle filter, nonlinear filter assigns weight each value value. Comparison final results proposed classification multilayer perceptron neural network, decision tree, simple Bayes, temporal fuzzy min–max network (TFMM-PSO) shows superiority method over comparable ones.