چاپ صفحه |  صفحه نخست سامانه |  نویسندگان |  اطلاعات تفضیلی |  دانلود مقاله | دانشگاه علوم پزشکی تبریز |
| نویسنده | نفر چندم مقاله |
|---|---|
| آرمان طلعت | اول |
| توحید جعفری کشکی | دوم |
| عنوان | متن |
|---|---|
| خلاصه مقاله | Cardiovascular disease (CVD) remains a leading cause of mortality worldwide, necessitating advanced predictive modeling techniques. This study investigates the performance of several machine learning survival analysis models – Cox-Time, Random Survival Forests (RSF), and Gradient Boosting Machines (GBM) – in comparison to an enhanced Cox Proportional Hazards (CPH) model incorporating non-linear effects and interactions. Utilizing data from the Azar cohort (n=15,001), we evaluated model performance using the time-dependent concordance index (TD-C index) and Integrated Brier Score (IBS). Results showed that Cox-Time and GBM achieved the highest mean TD-C index (0.75), indicating superior discriminatory power, while RSF demonstrated the lowest IBS (0.14), reflecting the best calibration. These findings underscore the potential of advanced machine learning models to capture complex relationships in survival data, especially where the proportional hazards assumption is violated. While offering enhanced predictive capabilities, the computational complexity and interpretability of these methods remain critical considerations for broader adoption in clinical settings. This study provides insights into the strengths and limitations of advanced machine learning techniques for predicting CVD outcomes. |
| کلمات کلیدی |
| نام فایل | تاریخ درج فایل | اندازه فایل | دانلود |
|---|---|---|---|
| 211.pdf | 1404/09/01 | 663984 | دانلود |
| 222.docx | 1404/09/01 | 57547 | دانلود |
