EUVIMEDEUVIMED entdecken
Crossref

redictive Modeling for Diabetic Complications Using Artificial Intelligence

Mohamed Elzain Gurashi Elshekh, Rami Yousef

Midocean Journal for Research and Studies · 2026

DOI: 10.71335/19ax5348 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Midocean Journal for Research and Studies, 2026. Verfasst von Mohamed Elzain Gurashi Elshekh und 1 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objective

To investigate the ability of machine learning to predict the development of cardiovascular complications (heart failure and myocardial infarction) and renal complications (nephropathy and chronic kidney disease) among people with diabetes mellitus and thus detect those at risk earlier. Methodology: Data from the UCI "Diabetes 130-US Hospitals" dataset was used. The number of encounters included in the dataset is more than 100,000. ICD-9 diagnostic codes were utilized to define cardiac and renal complications. Prior to building models, the data was prepared, cleaned, and features were engineered. XGBoost, CatBoost, and Random Forest machine learning models were developed and validated. Also, SHAP and LIME analyses were performed to make the process more transparent.

Results

The best-performing algorithm proved to be CatBoost, which obtained an AUC score of 0.94 in case of cardiac complications and 0.79 in case of renal outcomes. The other algorithms performed relatively well as well. The predictors identified with the help of SHAP and LIME analysis turned out to be consistent with some known clinical risk factors.

Conclusion

The presented results show that there is a lot of potential in using machine learning techniques for predicting cardiovascular and renal complications among people with diabetes. Machine learning models could help in managing patients proactively and could serve as a tool for doctors' decision making.

Zitieren

APA

Elshekh, M. E. G., & Yousef, R. (2026). redictive Modeling for Diabetic Complications Using Artificial Intelligence. Midocean Journal for Research and Studies. https://doi.org/10.71335/19ax5348

Vancouver

Elshekh MEG, Yousef R. redictive Modeling for Diabetic Complications Using Artificial Intelligence. Midocean Journal for Research and Studies. 2026. doi:10.71335/19ax5348

BibTeX
@article{Elshekh2026,
  author = {Elshekh, Mohamed Elzain Gurashi and Yousef, Rami},
  title = {redictive Modeling for Diabetic Complications Using Artificial Intelligence},
  journal = {Midocean Journal for Research and Studies},
  year = {2026},
  doi = {10.71335/19ax5348}
}
Kostenlos starten →