The cenreg Package

The Python package cenreg is a repository for probabilistic forecasts such as quantile regression and distribution regression and for censored regression such as survival analysis and interval-censored data analysis.

Features:

• Tree-based models run fast and output accurate predictions.
• Neural network models are implemented both for structured data (e.g., tabular data) and non-structured data (e.g., image data).
• Both models can handle competing risks.
• Both models are based on the (conditional) independence assumption or the non-informative assuption, but they can also handle dependent censoring based on assumed copula.
• Strictly proper scoring rules are implemented to evaluate the discrimination performances of prediction models. The scoring rules can handle right-censored and interval-censored data.
• Binning-free calibration metrics are implemented to evaluate the calibration performances of prediction models. The calibration metrics can handle right-censored and interval-censored data.

Getting Started

Prerequisites

You first need to install SurvSet via pip

pip install SurvSet

Additionally, denpending on the models you want to use, you also need to install

• LightGBM
• PyTorch

Installation

You can install cenreg via pip:

pip install cenreg

Run Sample Code

You can find our sample codes in the notebooks directory.

Documentation

Read the documentation to get started.

Citation

• H. Yanagisawa and S. Akiyama, A Strictly Proper Scoring Rule and a Calibration Metric for Interval-Censored Data Analysis, ICML 2026 (Paper in OpenReview)
• H. Yanagisawa and S. Akiyama, Survival Analysis via Density Estimation, ICML 2025 (Paper in OpenReview)
• H. Yanagisawa, Proper Scoring Rules for Survival Analysis, ICML 2023