PuSH - Publikationsserver des Helmholtz Zentrums München

Rügamer, D.* ; Kolb, C.* ; Fritz, C.* ; Pfisterer, F.* ; Kopper, P.* ; Bischl, B.* ; Shen, R.* ; Bukas, C. ; Barros De Andrade E Sousa, L. ; Thalmeier, D. ; Baumann, P.F.M.* ; Kook, L.* ; Klein, N.* ; Müller, C.L.

deepregression: A flexible neural network framework for semi-structured deep distributional regression.

J. Stat. Software 105, 1-31 (2023)
Postprint DOI
In this paper we describe the implementation of semi-structured deep distributional re-gression, a flexible framework to learn conditional distributions based on the combination of additive regression models and deep networks. Our implementation encompasses (1) a modular neural network building system based on the deep learning library TensorFlow for the fusion of various statistical and deep learning approaches, (2) an orthogonaliza-tion cell to allow for an interpretable combination of different subnetworks, as well as (3) pre-processing steps necessary to set up such models. The software package allows to define models in a user-friendly manner via a formula interface that is inspired by classical statistical model frameworks such as mgcv. The package’s modular design and function-ality provides a unique resource for both scalable estimation of complex statistical models and the combination of approaches from deep learning and statistics. This allows for state-of-the-art predictive performance while simultaneously retaining the indispensable interpretability of classical statistical models.
Altmetric
Weitere Metriken?
[➜Einloggen]
Zusatzinfos bearbeiten [➜Einloggen]
Publikationstyp Artikel: Journalartikel
Dokumenttyp Wissenschaftlicher Artikel
Korrespondenzautor
Schlagwörter Additive Predictors ; Deep Learning ; Effect Decomposition ; Orthogonal Complement ; Penalization ; Smoothing
ISSN (print) / ISBN 1548-7660
Zeitschrift Journal of Statistical Software
Quellenangaben Band: 105, Heft: 2, Seiten: 1-31 Artikelnummer: , Supplement: ,
Verlag University of California at Los Angeles
Verlagsort Ucla Dept Statistics, 8130 Math Sciences Bldg, Box 951554, Los Angeles, Ca 90095-1554 Usa
Nichtpatentliteratur Publikationen
Begutachtungsstatus Peer reviewed
Institut(e) Institute of Computational Biology (ICB)
Helmholtz Artifical Intelligence Cooperation Unit (HAICU)
Förderungen German research foundation (DFG) through the Emmy Noether grant
German Federal Ministry of Education and Research (BMBF)