Diagnosis of acute hyperglycemia based on data-driven prediction models

Area of Science:

• Endocrinology and Metabolic Disorders
• Medical Informatics
• Machine Learning in Healthcare

Background:

• Acute hyperglycemia is a critical endocrine and metabolic disorder with significant health implications.
• Effective diagnostic methods are crucial for improving treatment quality and patient satisfaction in acute hyperglycemia.
• Current research hotspots focus on developing advanced strategies for managing this condition.

Purpose of the Study:

• To introduce a novel data-driven prediction model for the diagnosis of acute hyperglycemia.
• To evaluate the efficacy of a Support Vector Machine (SVM) model in predicting acute hyperglycemia.
• To establish a clinical auxiliary diagnostic tool for early detection and treatment of acute hyperglycemia.

Main Methods:

• Collected clinical data from 1000 patients diagnosed with acute hyperglycemia.
• Performed data cleaning and feature engineering to select 10 relevant features, including BMI, TG, and HDL-C.
• Utilized a Support Vector Machine (SVM) model for training and testing the prediction system.

Main Results:

• The SVM model demonstrated high predictive performance for acute hyperglycemia.
• Achieved an average accuracy of 96%, a recall rate of 84%, and an F1 score of 89%.
• The data-driven diagnostic method proved effective in identifying patients with acute hyperglycemia.

Conclusions:

• Data-driven prediction models, specifically SVM, offer a valuable approach for diagnosing acute hyperglycemia.
• The developed method can serve as an effective clinical auxiliary diagnostic tool.
• Early diagnosis and treatment success rates for acute hyperglycemia can be improved using this predictive model.