Time Series Treatment Effects Analysis with Always-Missing Controls

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shu, Juan, Han, Qiyu, Chen, George, Cao, Xihao, Luo, Kangming, Pallotta, Dan, Agrawal, Shivam, Lu, Yuping, Zhang, Xiaoyu, Mansoor, Jawad, Anand, Jyoti
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909499465400320
author Shu, Juan
Han, Qiyu
Chen, George
Cao, Xihao
Luo, Kangming
Pallotta, Dan
Agrawal, Shivam
Lu, Yuping
Zhang, Xiaoyu
Mansoor, Jawad
Anand, Jyoti
author_facet Shu, Juan
Han, Qiyu
Chen, George
Cao, Xihao
Luo, Kangming
Pallotta, Dan
Agrawal, Shivam
Lu, Yuping
Zhang, Xiaoyu
Mansoor, Jawad
Anand, Jyoti
contents Estimating treatment effects in time series data presents a significant challenge, especially when the control group is always unobservable. For example, in analyzing the effects of Christmas on retail sales, we lack direct observation of what would have occurred in late December without the Christmas impact. To address this, we try to recover the control group in the event period while accounting for confounders and temporal dependencies. Experimental results on the M5 Walmart retail sales data demonstrate robust estimation of the potential outcome of the control group as well as accurate predicted holiday effect. Furthermore, we provided theoretical guarantees for the estimated treatment effect, proving its consistency and asymptotic normality. The proposed methodology is applicable not only to this always-missing control scenario but also in other conventional time series causal inference settings.
format Preprint
id arxiv_https___arxiv_org_abs_2502_12393
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Time Series Treatment Effects Analysis with Always-Missing Controls
Shu, Juan
Han, Qiyu
Chen, George
Cao, Xihao
Luo, Kangming
Pallotta, Dan
Agrawal, Shivam
Lu, Yuping
Zhang, Xiaoyu
Mansoor, Jawad
Anand, Jyoti
Methodology
Artificial Intelligence
Machine Learning
Estimating treatment effects in time series data presents a significant challenge, especially when the control group is always unobservable. For example, in analyzing the effects of Christmas on retail sales, we lack direct observation of what would have occurred in late December without the Christmas impact. To address this, we try to recover the control group in the event period while accounting for confounders and temporal dependencies. Experimental results on the M5 Walmart retail sales data demonstrate robust estimation of the potential outcome of the control group as well as accurate predicted holiday effect. Furthermore, we provided theoretical guarantees for the estimated treatment effect, proving its consistency and asymptotic normality. The proposed methodology is applicable not only to this always-missing control scenario but also in other conventional time series causal inference settings.
title Time Series Treatment Effects Analysis with Always-Missing Controls
topic Methodology
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2502.12393