A novel swarm-intelligent based gated recurrent unit for classifying and diagnosing diabetes mellitus patients

Abstract

Diabetes mellitus (DM) is a prevalent and complicated chronic condition that poses significant challenges to accurate diagnosis and classification, both of which are essential for innovative ways to enhance patient treatment. Diabetes has several negative effects, including heart disease, neuropathy, renal issues, and visual issues. We introduce a new and efficient method in this study, the Dynamic Glowworm Swarm Optimization Fine-Tuned Gated Recurrent Unit (DGSO-FTGRU), for the diagnosis and classification of DM patients. After gathering the Pima Indian dataset, the data was preprocessed using sigmoid normalization. Sidmoid normalization is a helpful method for preparing patient data for diabetics. It helps create accurate and comprehensible models for diagnosis, risk assessment, and treatment planning. Finding significant qualities that influence diabetes is made easier with the use of the DGSO-FTGRU, which provides valuable information on the importance of different features. By enabling quicker and more personalized interventions, this innovative technique delivers considerable advances in the control of diabetes and enhances patient health. The results of the study examined a number of variables, including recall (96.32%), accuracy (95.5%), specificity (92.43%), f-measures (95.41%), and precision (97.28%). In addition to providing practical medical solutions, the DGSO-FTGRU safeguards privacy laws and encourages patient data protection and confidentiality.