User: Guest  Login
Document type:
Zeitungsartikel 
Author(s):
Gissibl, N., Klüppelberg, C. and Lauritzen, S. 
Title:
Identifiability and estimation of recursive max-linear models 
Abstract:
We address the identifiability and estimation of recursive max-linear structural equation models represented by an edge weighted directed acyclic graph (DAG). Such models are generally unidentifiable and we identify the whole class of DAGs and edge weights corresponding to a given observational distribution. For estimation, standard likelihood theory cannot be applied because the corresponding families of distributions are not dominated. Given the underlying DAG, we present an estimator for the...    »
 
Keywords:
Causal inference, Bayesian network, directed acyclic graph, extreme value theory, generalized maximum likelihood estimation, graphical model, identifiability, max-linear model, structural equation model. 
Dewey Decimal Classification:
510 Mathematik 
Journal title:
Scandinavian Journal of Statistics 
Year:
2020 
Year / month:
2020-04 
Quarter:
2. Quartal 
Month:
Apr 
Language:
en 
Fulltext / DOI:
WWW:
_blank 
Publisher:
Wiley 
Notes:
online Veröffentlichung: 28. April 2020 
Status:
Erstveröffentlichung 
TUM Institution:
Lehrstuhl für Mathematische Statistik 
Format:
Text