Robust quantile regression in RKHS: Solution paths for censored and truncated data

초록

This paper proposes a fast and efficient algorithm for estimating quantile regression functions in a Reproducing Kernel Hilbert Space (RKHS) for data subject to left truncation and right censoring. By integrating a weighted estimation scheme for left-truncated and right-censored data with a solution path algorithm for the epsilon-insensitive loss, we develop a method that computes the entire path of quantile estimates as the regularization parameter varies. This approach avoids the computational cost of grid-search-based model selection and provides a robust tool for survival analysis in the presence of non-homogeneous variability and complex covariate relationships.

키워드