Additional file 1 of Multi-institutional development and external validation of machine learning-based models to predict relapse risk of pancreatic ductal adenocarcinoma after radical resection

Autor: Li, Xiawei, Yang, Litao, Yuan, Zheping, Lou, Jianyao, Fan, Yiqun, Shi, Aiguang, Huang, Junjie, Zhao, Mingchen, Wu, Yulian
Rok vydání: 2021
DOI: 10.6084/m9.figshare.14889172
Popis: Additional file 1: Figure S1. Comparisons of AUROC of different models to predict 1- and 2-year relapse in training set (a: 1-year relapse, b: 2-year relapse). Figure S2. Calibration curves of SVM model (A: training set; B: validation set) to predict 1-year relapse and KNN model (C: training set; D: validation set) to predict 2-year relapse. Table S1. The ranges of training parameters for grid search in different models. Table S2. Comparison of characteristics between patients with and without 1-year relapse in training set. Table S3. Comparison of characteristics between patients with and without 2-year relapse in training set. Table S4. The optimal parameters for different models to predict relapse risks of PDAC. Table S5. Performance of models built on all 32 variables in the validation set. Table S6. Performance of models built on variables from lasso analysis in the validation set.
Databáze: OpenAIRE