Causal Effect Identification in LiNGAM Models with Latent Confounders

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Tramontano, Daniele, Kivva, Yaroslav, Salehkaleybar, Saber, Drton, Mathias, Kiyavash, Negar
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914822038224896
author Tramontano, Daniele
Kivva, Yaroslav
Salehkaleybar, Saber
Drton, Mathias
Kiyavash, Negar
author_facet Tramontano, Daniele
Kivva, Yaroslav
Salehkaleybar, Saber
Drton, Mathias
Kiyavash, Negar
contents We study the generic identifiability of causal effects in linear non-Gaussian acyclic models (LiNGAM) with latent variables. We consider the problem in two main settings: When the causal graph is known a priori, and when it is unknown. In both settings, we provide a complete graphical characterization of the identifiable direct or total causal effects among observed variables. Moreover, we propose efficient algorithms to certify the graphical conditions. Finally, we propose an adaptation of the reconstruction independent component analysis (RICA) algorithm that estimates the causal effects from the observational data given the causal graph. Experimental results show the effectiveness of the proposed method in estimating the causal effects.
format Preprint
id arxiv_https___arxiv_org_abs_2406_02049
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Causal Effect Identification in LiNGAM Models with Latent Confounders
Tramontano, Daniele
Kivva, Yaroslav
Salehkaleybar, Saber
Drton, Mathias
Kiyavash, Negar
Machine Learning
Methodology
We study the generic identifiability of causal effects in linear non-Gaussian acyclic models (LiNGAM) with latent variables. We consider the problem in two main settings: When the causal graph is known a priori, and when it is unknown. In both settings, we provide a complete graphical characterization of the identifiable direct or total causal effects among observed variables. Moreover, we propose efficient algorithms to certify the graphical conditions. Finally, we propose an adaptation of the reconstruction independent component analysis (RICA) algorithm that estimates the causal effects from the observational data given the causal graph. Experimental results show the effectiveness of the proposed method in estimating the causal effects.
title Causal Effect Identification in LiNGAM Models with Latent Confounders
topic Machine Learning
Methodology
url https://arxiv.org/abs/2406.02049