Machine Learning-based Risk Assessment for Early Screening of Elderly Patients with Hyperglycemia

Abstract:
Background: In recent years, machine learning has become a promising approach for disease prediction. The increasing risk of hyperuricemia makes early screening for this condition necessary. Objective: This study aims to propose a high-performance framework for accurate prediction of hyperuricemia risk. Methods: Elderly health examination data from 2019, 2022, and 2023 from Tangqiao Community Hospital in Shanghai was selected. After data cleaning, important features were screened, and nine key indicators were chosen for risk analysis and prediction using the Extreme Gradient Boosting model. Additionally, the model's performance was validated by comparing it to three traditional machine learning models (decision tree, random forest, logistic regression). Results: Among the four different models selected, Extreme Gradient Boosting demonstrated relatively optimal performance, with an accuracy of 99%, precision of 99%, recall of 99%, F1 score of 99%, and area under the curve (AUC) of 1.0. Conclusion: This study proposes a hyperuricemia risk prediction model that achieves accurate prediction of hyperuricemia risk, contributing to the development of medical undertakings.