We address the problem of constraint-based causal discovery with mixed data types, such as (but not limited to) continuous, binary, multinomial and or-dinal variables. We use likelihood-ratio tests based on appropriate regression models, and show how to derive symmetric conditional independence tests. Such tests can then be directly used by existing constraint-based methods with mixed data, such as the PC and FCI algorithms for learning Bayesian networks and maximal ancestral graphs respectively. In experiments on simu-lated Bayesian networks, we employ the PC algorithm with different conditional independence tests for mixed data, and show that the proposed approach outperforms alternatives in terms of learning accuracy.
%0 Conference Paper
%1 Tsagris2017
%A Tsagris, Michail
%A Borboudakis, Giorgos
%A Lagani, Vincenzo
%A Tsamardinos, Ioannis
%B 23d ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2017)
%D 2017
%K Bayesian ancestral data graphs independence learning networks tests,Constraint-based {\textperiodcentered} {\textperiodcentered},Conditional {\textperiodcentered},Maximal {\textperiodcentered},Mixed
%T Constraint-based Causal Discovery with Mixed Data
%X We address the problem of constraint-based causal discovery with mixed data types, such as (but not limited to) continuous, binary, multinomial and or-dinal variables. We use likelihood-ratio tests based on appropriate regression models, and show how to derive symmetric conditional independence tests. Such tests can then be directly used by existing constraint-based methods with mixed data, such as the PC and FCI algorithms for learning Bayesian networks and maximal ancestral graphs respectively. In experiments on simu-lated Bayesian networks, we employ the PC algorithm with different conditional independence tests for mixed data, and show that the proposed approach outperforms alternatives in terms of learning accuracy.
@inproceedings{Tsagris2017,
abstract = {We address the problem of constraint-based causal discovery with mixed data types, such as (but not limited to) continuous, binary, multinomial and or-dinal variables. We use likelihood-ratio tests based on appropriate regression models, and show how to derive symmetric conditional independence tests. Such tests can then be directly used by existing constraint-based methods with mixed data, such as the PC and FCI algorithms for learning Bayesian networks and maximal ancestral graphs respectively. In experiments on simu-lated Bayesian networks, we employ the PC algorithm with different conditional independence tests for mixed data, and show that the proposed approach outperforms alternatives in terms of learning accuracy.},
added-at = {2018-11-08T11:26:15.000+0100},
author = {Tsagris, Michail and Borboudakis, Giorgos and Lagani, Vincenzo and Tsamardinos, Ioannis},
biburl = {https://www.bibsonomy.org/bibtex/20adffed04545697f97b75f30bf551834/nikolaof},
booktitle = {23d ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2017)},
file = {:C$\backslash$:/Users/MxM/AppData/Local/Mendeley Ltd./Mendeley Desktop/Downloaded/Tsagris et al. - 2017 - Constraint-based Causal Discovery with Mixed Data.pdf:pdf},
interhash = {87d6a33d891429260e644392ddcba508},
intrahash = {0adffed04545697f97b75f30bf551834},
keywords = {Bayesian ancestral data graphs independence learning networks tests,Constraint-based {\textperiodcentered} {\textperiodcentered},Conditional {\textperiodcentered},Maximal {\textperiodcentered},Mixed},
timestamp = {2018-11-08T11:26:15.000+0100},
title = {{Constraint-based Causal Discovery with Mixed Data}},
year = 2017
}