Diabetes Prediction and Insulin Dosage Prediction Using Machine Learning

EasyChair Preprint no. 8504

Abstract

Chronic metabolic problem caused by the disease known as diabetes mellitus. Blood glucose levels (BGLs) need to be adjusted so that diabetes individuals may continue to lead normal lives without developing the major problems that have historically plagued those with the disease. However, for various reasons, most diabetes individuals do not have their blood glucose levels under good control. Traditional methods of prevention, such as a good diet and regular exercise, are essential for diabetes patients, but the correct insulin dose is the most critical factor in managing their blood glucose levels. In this study, we offer a model for determining the optimal insulin dosage for a diabetic patient that is based on gradient boosting. Patients' weight, fasting glucose, and gender were among the numerous variables used to train and evaluate the suggested model. Good results were obtained when using the suggested approach to estimate the optimal insulin dose.

Keyphrases: Artificial Neural Networks, data preprocessing, Diabetes, Gradient Boosting, machine learning, prediction, Regression, visualization

@Booklet{EasyChair:8504,
  author = {Saketh Madabushi and Madhuri Kesari and Shiva Teja Dusa and Arun Kumar Gongati and Venkata Rao Yanamadni},
  title = {Diabetes Prediction and Insulin Dosage Prediction Using Machine Learning},
  howpublished = {EasyChair Preprint no. 8504},

  year = {EasyChair, 2022}}