Title: Predicting diabetes using Cohen's Kappa blending ensemble learning

Authors: Isaac Kofi Nti; Owusu Nyarko-Boateng; Adebayo Felix Adekoya; Benjamin Asubam Weyori; Henrietta Pokuaa Adjei

Addresses: Department of Computer Science and Informatics, University of Energy and Natural Resources, Sunyani, Ghana; School of Information Technology, University of Cincinnati, OH, USA ' Department of Computer Science and Informatics, University of Energy and Natural Resources, Sunyani, Ghana ' Department of Computer Science and Informatics, University of Energy and Natural Resources, Sunyani, Ghana ' Department of Computer Science and Informatics, University of Energy and Natural Resources, Sunyani, Ghana ' Department of Computer Science, Sunyani Technical University, Sunyani, Ghana

Abstract: Diabetes is a well-known risk factor for early mortality and disability. As signatories to the 2030 Agenda for Sustainable Development, Member States set an ambitious objective of a one-third reduction in early death due to non-communicable diseases (NCDs), which includes diabetes. Nonetheless, the current economic impact of diabetes on countries, individuals, and healthcare requires an agent means of its early detection. However, early detection of diabetes with conventional techniques is a considerable challenge for the healthcare industry and physicians. This study proposed a blended ensemble predictive model with Cohen's Kappa correlation-based base-learners selection to decrease unnecessary diabetes-related mortality through early detection. The empirical outcome shows that our proposed predictive model outperformed existing state-of-the-art approaches for predicting diabetes, thus resulting in enhanced diabetes prediction ability.

Keywords: diabetes; blended ensemble; diabetes prediction; Kappa statistic.

DOI: 10.1504/IJEH.2023.128605

International Journal of Electronic Healthcare, 2023 Vol.13 No.1, pp.57 - 70

Received: 07 Feb 2022
Accepted: 28 Oct 2022
Published online: 27 Jan 2023 *

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