Urinary Bladder Inflammation Prediction with the Gray Wolf Optimization Algorithm and Multi-Layer Perceptron-Based Hybrid Architecture

Authors : Mehmet Akif Bülbül

Pages : 1185-1194

Doi:10.17798/bitlisfen.1360049

View : 56 | Download : 52

Publication Date : 2023-12-28

Article Type : Research

Abstract :In this study, a decision support system for bladder inflammation prediction is presented. The proposed decision support system is built by establishing a hybrid architecture with Gray wolf optimization algorithm (GWO) and Multi-layer perceptron (MLP) networks. In addition to optimizing the hyperparameters in the MLP structure with GWO, the hybrid architecture also optimizes the order of input values to be presented to the MLP structure. The Acute Inflammations data set in the UCI Machine Learning repository was used as the data set in the study. Classification operations were carried out on this data set with the models obtained with hybrid architecture, Decision trees, k-Nearest Neighbors and Support Vector Machines methods. The controversial findings presented as a result of experimental studies have shown that the proposed hybrid architecture produces more successful results than other machine learning methods used in the study. In addition, the MLP network structure optimized with the hybrid architecture offers a new diagnostic method in terms of patient decision support systems.
Keywords : Gray Wolf Optimization Algorithm, Multi-Layer Perceptron, Decision Support Systems.