ARCHIVES
VOL. 2, ISSUE 4 (2026)
Predictive modeling of diabetes using ensemble machine learning and patient health indicators
Authors
Marek Pawlicki
Abstract
Appropriate management to reduce the risk of
long-term complications. The increasing availability of patient health data has
created opportunities for machine learning techniques to support diabetes risk
prediction. This article presents a predictive modeling framework based on
ensemble machine learning and routinely collected patient health indicators.
The proposed framework focuses on combining complementary machine learning
algorithms, particularly Random Forest, XGBoost, and Gradient Boosting, to
identify patterns associated with diabetes. Relevant patient indicators may
include age, body mass index (BMI), glucose level, blood pressure, insulin,
pregnancy history, and other routinely collected clinical characteristics.
Previous research has demonstrated that machine learning can identify useful
patterns in diabetes-related health data, while ensemble methods can provide
robust predictive performance across different datasets (Chang et al.,
2022; Mohsen et al., 2023) [2, 6]. The proposed framework
incorporates data preprocessing, feature selection, class-imbalance management,
model training, ensemble construction, and performance evaluation using
accuracy, precision, recall, F1-score, and area under the receiver operating
characteristic curve (AUC). Explainable artificial intelligence techniques such
as SHAP can additionally be used to identify the health indicators that
contribute most strongly to model predictions. The framework can provide a
foundation for developing data-driven decision-support systems for early
diabetes risk assessment, although external validation and clinical evaluation
are required before practical deployment.
Download
Pages:25-28
How to cite this article:
Marek Pawlicki "Predictive modeling of diabetes using ensemble machine learning and patient health indicators". World Journal of Advanced Research
, Vol 2, Issue 4, 2026, Pages 25-28
Pdf, Certificate, Cover Page, Index Page, Editorial Page, Accept Letter
Please enter the email address corresponding to this article submission.
