Title: Early prediction of diabetes mellitus using various artificial intelligence techniques: a technological review
Authors: Shahid Mohammad Ganie; Majid Bashir Malik; Tasleem Arif
Addresses: Department of Computer Sciences, BGSB University, Rajouri, J&K, India ' Department of Computer Sciences, BGSB University, Rajouri, J&K, India ' Department of Information Technology, BGSB University, India
Abstract: Millions of people around the globe are suffering from diabetes. Most of the patients (diabetic or potentially diabetic) are not familiar with their health issues and the risk factor they face before the diagnosis of diabetes. The paper reviews substantial work related to diabetes mellitus based on different classification techniques. In this paper, a generic smart framework for realistic health management of diabetes mellitus is presented and implemented using a publically available Pima Indian diabetes dataset sourced from the UCI machine learning repository. Different classification algorithms were employed namely decision tree (DT), random forest (RF), eXtreme gradient boosting (XGB), AdaBoost (AB) and gradient boosting classifier (GBC). Pre-processing techniques have been employed to improve the data quality assessment. Among all the classifiers, GB outperformed other models with accuracy rate of 92.20% followed by RF, XGB, ADB and DT as 91.55%, 89.61%, 89.61% and 88.96%, respectively.
Keywords: diabetes mellitus; data mining; artificial intelligence; machine learning; deep learning; healthcare framework; Jupyter notebook; Python.
International Journal of Business Intelligence and Systems Engineering, 2021 Vol.1 No.4, pp.325 - 346
Received: 28 Nov 2020
Accepted: 02 Jun 2021
Published online: 10 May 2022 *