International Journal of Clinical Case Reports, 2025, Vol.15, No.5, 209-218 http://medscipublisher.com/index.php/ijccr 212 Figure 1 The framework of the MTL-Cox model (Adopted from Zhang et al., 2023) Image caption: (A) The inner circle shows the main sites of the nine chronic diseases, and the outer circle demonstrates factors that affect chronic diseases in the UK Biobank data, which is detailed in the Materials Section; The figure is created with Biorender; (B) The flow of the MTL-Cox model construction and optimization, which detailed in the Model Section; (C) Concordance index, AUC, specificity, sensitivity, and Youden index are used to evaluate models, which detailed in the Experiments and results Section; (D) Applying the MTL-Cox model for chronic diseases personalized prediction, which detailed in the experiments and results section (Adopted from Zhang et al., 2023) These traditional methods have some deficiencies when analyzing complex associations and many variables. When dealing with multi-source data or when the interaction between variables is very strong, the effect is usually not ideal. Recent studies have combined the Cox model with multi-task learning, which can simultaneously predict multiple diseases and utilize the information exchange between different tasks to improve the accuracy of individualized risk prediction. 4.2 Machine learning models: decision tree, random forest, SVM, XGBoost Machine learning methods are increasingly useful in the prediction of chronic diseases, and they are good at handling complex nonlinear connections. Decision trees can provide a simple and clear classification method, while combined methods like random forests and XGBoost, by integrating multiple decision trees, make predictions more stable and accurate. Support Vector Machine (SVM) also performs well when processing high-dimensional data and has a good effect in the analysis of various chronic disease data (Al-Jamimi, 2024). Studies have found that ensemble methods, especially gradient enhancement and random forest, are usually more accurate than traditional statistical methods when predicting a large number of different data (Khalid et al., 2023).
RkJQdWJsaXNoZXIy MjQ4ODYzNA==