Paper Key : IRJ************434
Author: Sirigireddy Bharath Chandra Reddy,Gudapati Sri Sri Ram Karthikeya,Kodi Nandan,Mullapudi Charan,Basaba Bikram
Date Published: 03 Jul 2024
Abstract
Diabetes mellitus, a chronic metabolic disorder characterized by elevated blood glucose levels, poses a significant global health challenge. Predictive modeling using machine learning techniques has emerged as a promising approach for early diagnosis and intervention. This study investigates the performance of various machine learning models in predicting diabetes based on comprehensive datasets. The dataset comprises a diverse set of features, including demographic information, clinical markers, and lifestyle factors. We employ logistic regression, decision trees, random forests, support vector machines, k-nearest neighbors, neural networks, and gradient boosting models (XGBoost, LightGBM) to assess their efficacy in diabetes prediction. The dataset undergoes rigorous preprocessing, including handling missing values and feature scaling, to ensure the robustness of the models. Comparative analysis reveals the strengths and limitations of each model, considering factors such as interpretability, accuracy, and computational efficiency. The results highlight the significance of feature selection and model tuning in optimizing predictive performance. Additionally, we discuss the implications of these findings for early diabetes detection and personalized healthcare. Our study contributes to the evolving landscape of diabetes prediction methodologies, offering insights into the suitability of different machine learning models for diverse datasets and laying the groundwork for future advancements in this critical field of medical research.
DOI Requested