Niklas Koenen

SEP 10, 2026 · PREPRINT

A Hilbert-Valued Functional Decomposition Framework for Explaining Time-Dependent Outputs

arXiv

This is a methodological framework paper presenting novel mathematical theory for feature explanations in functional outputs, validated on synthetic and real data but without clinical endpoints, comparative effectiveness, or evidence of impact on decision-making.

AUG 17, 2026 · PREPRINT

LiD-GLM: Lipschitz-constrained Deep Generalized Linear Models

arXiv

This is a methodological proposal for a hybrid statistical-neural model architecture with theoretical justification but no empirical validation against clinical or real-world outcomes reported in the source.