Title: Identification and prioritisation of diabetic nephropathy risk factors in diabetes patients using machine learning approach

Authors: Seyyed Mahdi Hosseini Sarkhosh; Mahdieh Taghvaei

Addresses: Department of Industrial Engineering, University of Garmsar, Garmsar, Iran ' Department of Internal Medicine, Pakdasht Hospital, Shahid Beheshti University of Medical Sciences, Tehran, Iran

Abstract: A significant microvascular complication in diabetic patients is diabetic nephropathy (DN). Despite the fact that different risk factors were previously identified for DN, machine learning (ML) techniques can confirm the importance of the predictive factors and determine their priority. Hence, this research is primarily aimed to identify and prioritise DN predictive risk factors among patients suffering from type 2 diabetes mellitus (T2DM) using the ML techniques. The characteristics of 2703 patients with T2DM are obtained from the dataset of the National Health and Nutrition Examination Survey (NHANES). Then, the recursive feature elimination using the cross-validation (RFECV) technique is used to select the essential factors. Next, five classification algorithms are utilised to construct the predictive model. The results confirm the known predictors for DN and emphasise the importance of controlling hypertension, lipid, weight, glucose, and uric acid in patients suffering from T2DM.

Keywords: diabetic nephropathy; machine learning; risk factors; NHANES.

DOI: 10.1504/IJDMB.2021.122862

International Journal of Data Mining and Bioinformatics, 2021 Vol.25 No.3/4, pp.216 - 233

Received: 04 Feb 2021
Accepted: 28 Feb 2022

Published online: 13 May 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article