# Censored Quantile Regression Forests

## TL;DR

This paper introduces Censored Quantile Regression Forests, a novel non-parametric method for estimating time-to-event quantiles in censored data, improving predictive accuracy without parametric assumptions.

## Contribution

It develops a new regression adjustment for censored data based on adaptive estimating equations, extending random forests to censored quantile regression.

## Key findings

- Demonstrates consistency under mild conditions
- Shows superior performance in numerical studies
- Enables quantile estimation without parametric models

## Abstract

Random forests are powerful non-parametric regression method but are severely limited in their usage in the presence of randomly censored observations, and naively applied can exhibit poor predictive performance due to the incurred biases. Based on a local adaptive representation of random forests, we develop its regression adjustment for randomly censored regression quantile models. Regression adjustment is based on new estimating equations that adapt to censoring and lead to quantile score whenever the data do not exhibit censoring. The proposed procedure named censored quantile regression forest, allows us to estimate quantiles of time-to-event without any parametric modeling assumption. We establish its consistency under mild model specifications. Numerical studies showcase a clear advantage of the proposed procedure.