article · Diagnostics
Diabetes mellitus is an epidemiological condition affecting populations globally, where machine learning methods can support diagnosis, assessment, and prognosis. A new computational approach introduces a hybrid metaheuristic optimisation algorithm that combines dynamic Al-Biruni earth radius and dipper-throated optimisation techniques. This system performs feature selection on relevant health indicators and simultaneously optimises the operating parameters of a random forest classification model. When benchmarked against alternative machine learning models and optimisation methods, the hybrid technique attained an overall diabetes classification accuracy of 98.6 per cent. Statistical evaluations using analysis of variance and Wilcoxon signed-rank tests confirmed that the method offers statistically significant performance advantages over the compared alternatives. The technique provides a refined method for identifying critical diagnostic features and improving algorithmic classification accuracy for chronic disease assessment.
Diabetes represents a major global health challenge requiring reliable diagnostic and prognostic tools. Utilising machine learning can help healthcare practitioners identify risk factors and detect conditions earlier. By improving both feature selection and model tuning, this approach increases algorithmic accuracy to over 98 per cent, potentially offering more dependable analytical foundations for future automated health screening systems.
This computational framework targets diagnostic software developers and digital health providers seeking higher precision in predictive analytics for chronic disease. Because testing remains confined to comparative algorithmic benchmarks and statistical evaluations against existing machine learning models, the technology represents early-stage research that requires validation within practical clinical workflows and regulatory pathways before real-world deployment can occur.
AI-generated from the published abstract. Always read the original work before citing.
INTRODUCTION: In public health, machine learning algorithms have been used to predict or diagnose chronic epidemiological disorders such as diabetes mellitus, which has reached epidemic proportions due to its widespread occurrence around the world. Diabetes is just one of several diseases for which machine learning techniques can be used in the diagnosis, prognosis, and assessment procedures. METHODOLOGY: In this paper, we propose a new approach for boosting the classification of diabetes based on a new metaheuristic optimization algorithm. The proposed approach proposes a new feature selection algorithm based on a dynamic Al-Biruni earth radius and dipper-throated optimization algorithm (DBERDTO). The selected features are then classified using a random forest classifier with its parameters optimized using the proposed DBERDTO. RESULTS: The proposed methodology is evaluated and compared with recent optimization methods and machine learning models to prove its efficiency and superiority. The overall accuracy of diabetes classification achieved by the proposed approach is 98.6%. On the other hand, statistical tests have been conducted to assess the significance and the statistical difference of the proposed approach based on the analysis of variance (ANOVA) and Wilcoxon signed-rank tests. CONCLUSIONS: The results of these tests confirmed the superiority of the proposed approach compared to the other classification and optimization methods.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.3390/diagnostics13122038
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
