Manuscript received December 25, 2023; revised May 30, 2024; accepted June 5, 2024; revised June 16, 2024
Abstract—Diabetes Mellitus (DM) is a condition in which the pancreas is incapable of producing enough insulin for glucose metabolism. Risk factors such as age, hectic schedules, inactivity, patient weight, high blood pressure, and blood sugar level are considered to be the primary cause of type 2 diabetes. Due to misinformation and bad eating habits, the pace of increase in diabetes individuals is problematic. Therefore, a framework employing clinical criteria to diagnose thousands of patients accurately is required. For predicting DM at an early stage based on the risk-based characteristics of a person's health, stacking-based classifier is developed that combines five classifiers, namely Logistic Regression (LR), AdaBoost + Support Vector Machine (SVM), Nave Bayes (NB), Artificial Neural Network (ANN), and k-Nearest Neighbors (k-NN), into a single model and uses Random Forest (RF) as a meta-learner. In addition, the performance of these six classifiers was compared to that of the stacked model using the PIMA Indians Diabetes Database (PIDD) dataset. The outcome of the performance analysis revealed that the proposed model obtained ~85.36% accuracy, which is much higher than the six classifiers.
Keywords—AdaBoost, logistic regression, Naïve Bayes, stacking
Cite: Aman and Rajender Singh Chhillar , " A Stacking-based Hybrid Model with Random Forest as Meta-learner for Diabetes Mellitus Prediction," International Journal of Machine Learning vol. 14, no. 2, pp. 54-58, 2024.
Copyright © 2024 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).