Deep and interpretable regression models for ordinal outcomes

Kook, Lucas; Herzog, Lisa; Hothorn, Torsten; Dürr, Oliver; Sick, Beate

doi:10.1016/j.patcog.2021.108263

Publication type:	Article in scientific journal
Type of review:	Peer review (publication)
Title:	Deep and interpretable regression models for ordinal outcomes
Authors:	Kook, Lucas Herzog, Lisa Hothorn, Torsten Dürr, Oliver Sick, Beate
et. al:	No
DOI:	10.1016/j.patcog.2021.108263
Published in:	Pattern Recognition
Volume(Issue):	122
Issue:	108263
Issue Date:	2022
Publisher / Ed. Institution:	Elsevier
ISSN:	0031-3203 1873-5142
Language:	English
Subjects:	Deep learning; Interpretability; Distributional regression; Ordinal regression; Transformation model
Subject (DDC):	006: Special computer methods
Abstract:	Outcomes with a natural order commonly occur in prediction problems and often the available input data are a mixture of complex data like images and tabular predictors. Deep Learning (DL) models are state-of-the-art for image classification tasks but frequently treat ordinal outcomes as unordered and lack interpretability. In contrast, classical ordinal regression models consider the outcome’s order and yield interpretable predictor effects but are limited to tabular data. We present ordinal neural network transformation models (ontrams), which unite DL with classical ordinal regression approaches. ontrams are a special case of transformation models and trade off flexibility and interpretability by additively decomposing the transformation function into terms for image and tabular data using jointly trained neural networks. The performance of the most flexible ontram is by definition equivalent to a standard multi-class DL model trained with cross-entropy while being faster in training when facing ordinal outcomes. Lastly, we discuss how to interpret model components for both tabular and image data on two publicly available datasets.
Further description:	A preprint version of this article is available on arXiv at https://doi.org/10.48550/arXiv.2010.08376
URI:	https://digitalcollection.zhaw.ch/handle/11475/27101
Fulltext version:	Published version
License (according to publishing contract):	Licence according to publishing contract
Departement:	School of Engineering
Organisational Unit:	Institute of Data Analysis and Process Design (IDP)
Appears in collections:	Publikationen School of Engineering

Files in This Item:

There are no files associated with this item.

Show full item record

Kook, L., Herzog, L., Hothorn, T., Dürr, O., & Sick, B. (2022). Deep and interpretable regression models for ordinal outcomes. Pattern Recognition, 122(108263). https://doi.org/10.1016/j.patcog.2021.108263

Kook, L. et al. (2022) ‘Deep and interpretable regression models for ordinal outcomes’, Pattern Recognition, 122(108263). Available at: https://doi.org/10.1016/j.patcog.2021.108263.

L. Kook, L. Herzog, T. Hothorn, O. Dürr, and B. Sick, “Deep and interpretable regression models for ordinal outcomes,” Pattern Recognition, vol. 122, no. 108263, 2022, doi: 10.1016/j.patcog.2021.108263.

KOOK, Lucas, Lisa HERZOG, Torsten HOTHORN, Oliver DÜRR und Beate SICK, 2022. Deep and interpretable regression models for ordinal outcomes. Pattern Recognition. 2022. Bd. 122, Nr. 108263. DOI 10.1016/j.patcog.2021.108263

Kook, Lucas, Lisa Herzog, Torsten Hothorn, Oliver Dürr, and Beate Sick. 2022. “Deep and Interpretable Regression Models for Ordinal Outcomes.” Pattern Recognition 122 (108263). https://doi.org/10.1016/j.patcog.2021.108263.

Kook, Lucas, et al. “Deep and Interpretable Regression Models for Ordinal Outcomes.” Pattern Recognition, vol. 122, no. 108263, 2022, https://doi.org/10.1016/j.patcog.2021.108263.