A Combined Method for Diabetes Mellitus Diagnosis Using Deep Learning, Singular Value Decomposition, and Self-Organizing Map Approaches

Aug 2, 2026

Diabetes is a rapidly growing chronic condition and a major societal health challenge, making accurate diabetic classification and risk assessment critical. Missing values in medical datasets can substantially degrade prediction performance. A combined machine-learning approach is proposed for diabetes mellitus diagnosis, integrating Singular Value Decomposition for missing value prediction, a Self-Organizing Map for clustering, STEPDISC for feature selection, and an ensemble of Deep Belief Network classifiers. Results on real-world datasets demonstrate accurate diabetes mellitus prediction compared with prior machine-learning methods.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages21

Reading Time53 min

AccessFree to download

Introduction

Diabetes mellitus background and risk factors

Machine learning for disease diagnosis

Proposed combined method

diagnostics Article A Combined Method for Diabetes Mellitus Diagnosis Using Deep Learning, Singular Value Decomposition, and Self-Organizing Map Approaches Mehrbakhsh Nilashi 1,2, *, Rabab Ali Abumalloh 3, Sultan Alyami 4, Abdullah Alghamdi 5 and Mesfer Alrizq 5 Citation: Nilashi, M.; Abumalloh, R.A.; Alyami, S.; Alghamdi, A.; Alrizq, M. A Combined Method for Diabetes Mellitus Diagnosis Using Deep Learning, Singular Value Decomposition, and Self-Organizing Map Approaches. Diagnostics 2023, 13, 1821. [https://doi.org/10.3390/](https://doi.org/10.3390/)[ ](https://doi.org/10.3390/)diagnostics13101821 Academic Editors: Sameer Antani and Dechang Chen Received: 9 February 2023 Revised: 10 March 2023 Accepted: 12 April 2023 Published: 22 May 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://](https://)[ ](https://)[creativecommons.org/licenses/by/](creativecommons.org/licenses/by/)[ ](creativecommons.org/licenses/by/)[4.0/](4.0/)) . 1 UCSI Graduate Business School, UCSI University, No. 1 Jalan Menara Gading, UCSI Heights, Cheras, Kuala Lumpur 56000, Malaysia 2 Centre for Global Sustainability Studies (CGSS), Universiti Sains Malaysia (USM), George Town 11800, Malaysia 3 Department of Computer Science and Engineering, Qatar University, Doha 2713, Qatar 4 Computer Science Department, College of Computer Science and Information Systems, Najran University, Najran 55461, Saudi Arabia 5 Information Systems Department, College of Computer Science and Information Systems, Najran University, Najran 55461, Saudi Arabia * [Correspondence: nilashidotnet@hotmail.com](Correspondence: nilashidotnet@hotmail.com) Abstract: Diabetes in humans is a rapidly expanding chronic disease and a major crisis in modern societies. The classiﬁcation of diabetics is a challenging and important procedure that allows the interpretation of diabetic data and diagnosis. Missing values in datasets can impact the prediction accuracy of the methods for the diagnosis. Due to this, a variety of machine learning techniques has been studied in the past. This research has developed a new method using machine learning techniques for diabetes risk prediction. The method was developed through the use of clustering and prediction learning techniques. The method uses Singular Value Decomposition for missing value predictions, a Self-Organizing Map for clustering the data, STEPDISC for feature selection, and an ensemble of Deep Belief Network classiﬁers for diabetes mellitus prediction. The performance of the proposed method is compared with the previous prediction methods developed by machine learning techniques. The results reveal that the deployed method can accurately predict diabetes mellitus fora set of real-world datasets. Keywords: Singular Value Decomposition; diabetes; Self-Organizing Map; diagnosis; accuracy; deep learning 1. Introduction Diabetes, also known as diabetes mellitus, in humans is a rapidly expanding chronic disease [1,2] and has major impacts on modern societies [3] . As a person with diabetes mellitus cannot process food properly, glucose builds up in the bloodstream [4,5] . With diabetes, the body cannot produce sufﬁcient insulin (type 1 diabetes) or cannot effectively utilize the hormone (type 2 diabetes) [6–8] . The body stops insulin in type 1 diabetes. In type 2 diabetes, the body's cells do not respond effectively to insulin. This is due to the body's immune system attacking and destroying a part of the pancreas mistakenly. Young people are usually affected by type 1 diabetes, mostly under 30 years of age. Type 2 diabetes, in contrast, often impacts middle-aged and older-aged people and is not fully curable. Some risk factors for type 2 diabetes are family history, age, ethnicity, being obese, having a diagnosis of gestational diabetes, high bl

FAQs

What problem does the combined method address in diabetes diagnosis?

It targets accurate diabetes mellitus classification and improves prediction when datasets contain missing values that can reduce accuracy.

How does Singular Value Decomposition (SVD) contribute to the method?

SVD is used for dimensionality reduction and missing value prediction within the overall pipeline.

What roles do Self-Organizing Map (SOM) and Deep Belief Network (DBN) play?

SOM performs data clustering, while an ensemble of Deep Belief Network classifiers is used for diabetes mellitus prediction after feature selection.

Aug 2, 2026

Upload by Emma Wilson