mltpy — Conditional Transformation Models#

Python port of Hothorn’s R mlt package. Fit flexible conditional distributions to continuous, censored, or covariate-dependent data using monotone Bernstein-polynomial transformations.

From a single fitted model, mltpy exposes the cumulative distribution, density, quantile, and hazard functions — and can simulate synthetic observations. The methodology is described in Hothorn, Kneib & Bühlmann (2014) and Hothorn (2020).

Getting started