stackBagg-package: stackBagg: Stacked IPCW Bagging

In pablogonzalezginestet/stackBagg: Stacked IPCW Bagging

stackBagg: Stacked IPCW Bagging

Description

Estimation the risk of an event at a specific time in the presence of censored data with competing risk using a collection of machine learning algorithms. Each algorithm is trained on a IPC-weighted bootstrap sample from the training set and predictions are obtained for the validation set. The stacked IPCW Bagging prediction is given by the weighted linear combination of the prediction of each algorithm. The area under the IPCW time-dependent receiver operator curve (AUC) is used for evaluating the predictive performance and calibrating the stack and each single algorithm.

Author(s)

Maintainer: Pablo Gonzalez Ginestet [e-mail] (https://staff.ki.se/people/pabgon)

pablogonzalezginestet/stackBagg documentation built on Aug. 27, 2023, 10:21 p.m.