Causal Matrix Completion under Multiple Treatments via Mixed Synthetic Nearest Neighbors

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Luo, Minrui, Zhang, Zhiheng
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918384383295488
author Luo, Minrui
Zhang, Zhiheng
author_facet Luo, Minrui
Zhang, Zhiheng
contents Synthetic Nearest Neighbors (SNN) provides a principled solution to causal matrix completion under missing-not-at-random (MNAR) by exploiting local low-rank structure through fully observed anchor submatrices. However, its effectiveness critically relies on sufficient data availability within each treatment level, a condition that often fails in settings with multiple or complex treatments. In this work, we propose Mixed Synthetic Nearest Neighbors (MSNN), a new entry-wise causal identification estimator that integrates information across treatment levels. We show that MSNN retains the finite-sample error bounds and asymptotic normality guarantees of SNN, while enlarging the effective sample size available for estimation. Empirical results on synthetic and real-world datasets illustrate the efficacy of the proposed approach, especially under data-scarce treatment levels.
format Preprint
id arxiv_https___arxiv_org_abs_2603_11942
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Causal Matrix Completion under Multiple Treatments via Mixed Synthetic Nearest Neighbors
Luo, Minrui
Zhang, Zhiheng
Machine Learning
Synthetic Nearest Neighbors (SNN) provides a principled solution to causal matrix completion under missing-not-at-random (MNAR) by exploiting local low-rank structure through fully observed anchor submatrices. However, its effectiveness critically relies on sufficient data availability within each treatment level, a condition that often fails in settings with multiple or complex treatments. In this work, we propose Mixed Synthetic Nearest Neighbors (MSNN), a new entry-wise causal identification estimator that integrates information across treatment levels. We show that MSNN retains the finite-sample error bounds and asymptotic normality guarantees of SNN, while enlarging the effective sample size available for estimation. Empirical results on synthetic and real-world datasets illustrate the efficacy of the proposed approach, especially under data-scarce treatment levels.
title Causal Matrix Completion under Multiple Treatments via Mixed Synthetic Nearest Neighbors
topic Machine Learning
url https://arxiv.org/abs/2603.11942