Salvato in:
Dettagli Bibliografici
Autori principali: Crasson, Thomas, Nabet, Yacine, Lécuyer, Mathias
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2411.00126
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909374384963584
author Crasson, Thomas
Nabet, Yacine
Lécuyer, Mathias
author_facet Crasson, Thomas
Nabet, Yacine
Lécuyer, Mathias
contents Deep learning time-series models are often used to make forecasts that inform downstream decisions. Since these decisions can differ from those in the training set, there is an implicit requirement that time-series models will generalize outside of their training distribution. Despite this core requirement, time-series models are typically trained and evaluated on in-distribution predictive tasks. We extend the orthogonal statistical learning framework to train causal time-series models that generalize better when forecasting the effect of actions outside of their training distribution. To evaluate these models, we leverage Regression Discontinuity Designs popular in economics to construct a test set of causal treatment effects.
format Preprint
id arxiv_https___arxiv_org_abs_2411_00126
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Training and Evaluating Causal Forecasting Models for Time-Series
Crasson, Thomas
Nabet, Yacine
Lécuyer, Mathias
Machine Learning
Artificial Intelligence
Deep learning time-series models are often used to make forecasts that inform downstream decisions. Since these decisions can differ from those in the training set, there is an implicit requirement that time-series models will generalize outside of their training distribution. Despite this core requirement, time-series models are typically trained and evaluated on in-distribution predictive tasks. We extend the orthogonal statistical learning framework to train causal time-series models that generalize better when forecasting the effect of actions outside of their training distribution. To evaluate these models, we leverage Regression Discontinuity Designs popular in economics to construct a test set of causal treatment effects.
title Training and Evaluating Causal Forecasting Models for Time-Series
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2411.00126