Prediction of recurrence of diffuse toxic goiter based on clinical data and comparison of machine learning methods and statistical models

This paper analyzes models for predicting clinical events using the example of recurrence of diffuse toxic goiter (DTG), employing machine learning methods and statistical approaches. The study includes data from 185 patients. To identify the most significant predictors of DTG recurrence, random forest and extreme gradient boosting algorithms were applied, demonstrating higher predictive performance (AUC 0.82–0.83) compared to traditional statistical models. The results of this study open new prospects for the development of personalized medicine, particularly in approaches to the treatment of patients with DTG.

Authors: E. M. Dariy, A. A. Meldo, S. V. Dora, Yu. Sh. Khalimov

Direction: Informatics, Computer Technologies And Control

Keywords: diffuse toxic goiter, machine learning, clinical event prediction, recurrence prediction, thyrotoxicosis, risk factors

---

View full article