cs.LGNov 9, 2026

Odds-Structured Distributional Regression for Interpretable Time-to-Event Prediction

Authors: Jakob Versnjak, Manuel Welte, Titus Kuehne, Marcus Kelm, Grégoire Montavon

Organizations: BIFOLD – Berlin Institute for the Foundations of Learning and Data, Berlin, Germany · Deutsches Herzzentrum der Charité, Institute of Computer-assisted Cardiovascular Medicine, Augustenburger Platz 1, 13353 Berlin, Germany · Charité – Universitätsmedizin Berlin, Corporate Member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Charitéplatz 1, Berlin 10117, Germany · Deutsches Herzzentrum der Charité, Institute for AI in Medicine, Invalidenstraße 120/121, 10115 Berlin, Germany · 

Abstract

Time-to-event prediction is fundamentally challenged by partial observability and aleatoric uncertainty. In clinical medicine, for instance, the timing of disease progression or acute readmission is contingent on highly dynamic and largely latent health trajectories. Rather than forcing a deterministic point forecast, distributional regression offers a principled means to address this complexity by predicting a probability distribution over time. However, explaining these distributional predictions presents a key obstacle: raw distribution parameters output by the model are abstract mathematical quantities that do not align well with domain concepts. To gain interpretability, we propose a novel paradigm where output parameters are designed to be additive in log-odds---a widely established measure in clinical risk modeling. Our approach, called ODDS-DR, moves beyond technical convenience to deliver clear domain alignment as well as the layer-wise computational structure required for robust, time-dependent feature attribution. Validated across two real-world clinical cohorts, our framework matches state-of-the-art predictive performance while resolving distributional outputs into interpretable risk markers across flexible prognostic horizons.

Figures & tables

Explore similar work

CardsList
  1. A Practical Guide Towards Interpreting Time-Series Deep Clinical Predictive Models: A Reproducibility Study

    Mar 25, 2026Yongda Fan, John Wu, Andrea Fitzpatrick +3Explainable Time Series ForecastingNeural Network Interpretability

  2. INTERVenE: Temporal-Abstraction-Interval Based Transformers for Short-Horizon Medical Event Prediction

    Aug 30, 2026Shahar Oded, Yuval ShaharEvent Time PredictionInterpretable ML

  3. TA-RNN-Medical-Hybrid: A Time-Aware and Interpretable Framework for Mortality Risk Prediction

    Mar 9, 2026Zahra Jafari, Azadeh Zamanifar, Amirfarhad FarhadiMortality PredictionLongitudinal EHR Modeling