Identifiability of total effects from abstractions of time series causal graphs - Centre Henri Lebesgue Accéder directement au contenu
Communication Dans Un Congrès Année : 2024

Identifiability of total effects from abstractions of time series causal graphs

Résumé

We study the problem of identifiability of the total effect of an intervention from observational time series in the situation, common in practice, where one only has access to abstractions of the true causal graph. We consider here two abstractions: the extended summary causal graph, which conflates all lagged causal relations but distinguishes between lagged and instantaneous relations, and the summary causal graph which does not give any indication about the lag between causal relations. We show that the total effect is always identifiable in extended summary causal graphs and provide sufficient conditions for identifiability in summary causal graphs. We furthermore provide adjustment sets allowing to estimate the total effect whenever it is identifiable.
Fichier principal
Vignette du fichier
Identification_Summary_UAI24_Camera_Ready.pdf (343.84 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04250602 , version 1 (19-10-2023)
hal-04250602 , version 2 (24-10-2023)
hal-04250602 , version 3 (20-02-2024)
hal-04250602 , version 4 (30-05-2024)

Licence

Identifiants

Citer

Charles Assaad, Emilie Devijver, Éric Gaussier, Gregor Gössler, Anouar Meynaoui. Identifiability of total effects from abstractions of time series causal graphs. 40th Conference on Uncertainty in Artificial Intelligence, Association for Uncertainty in Artificial Intelligence (AUAI), Jul 2024, Barcelone, Spain. ⟨hal-04250602v4⟩
98 Consultations
40 Téléchargements

Altmetric

Partager

Gmail Mastodon Facebook X LinkedIn More