Diabetes Prediction Using Nature-Inspired Optimization Algorithms

Abstract

Diabetes is a common genetic disease across the world. It is a chronic disease spurred by either production of insufficient production of insulin by the pancreas or less insulin use by the body. Insulin maintains the blood sugar level of the body. Uncontrolled diabetes can cause hypertension, also known as high blood glucose. Hypertension in the long run can be a cause of heart disease. Therefore, the prediction of diabetes is a major problem in the health sector. Various machine learning methods are already in use for the prediction of diabetes. Traditional methods were first used to address diabetes forecasting, but they take a while to deploy. Therefore, future research has adopted the use of various optimization strategies to solve the issue of diabetes forecasting as well as to boost system selectivity and operating speed. These methods have all proven to be successful in solving different optimization issues and also improve the model significantly. This work proposes a Logistic Regression model by the use of nature-inspired algorithms. We proposed a methodology which uses seven different optimization algorithms instead of gradient descent for linear regression. We implemented a linear regression algorithm from scratch and then used it with different optimization algorithms which improve the overall performance.