Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Integrating locally learned causal structures with overlapping variables
 

Summary: Integrating locally learned causal structures with
overlapping variables
Anonymous Author(s)
Affiliation
Address
email
Abstract
In many domains, data are distributed among datasets that share only some vari-
ables; other recorded variables may occur in only one dataset. There are several
asymptotically correct, informative algorithms that search for causal information
given a single dataset, even with missing values and hidden variables. There are,
however, no such reliable procedures for distributed data with overlapping vari-
ables, and only a single heuristic procedure (Structural EM). This paper describes
an asymptotically correct procedure, ION, that provides all the information about
structure obtainable from the marginal independence relations. Using simulated
and real data, the accuracy of ION is compared with that of Structural EM, and
with inference on complete, unified data.
1 Introduction
In many domains, researchers are interested in predicting the effects of variable interventions, such
as policy changes. Such predictions require knowledge of causal relations, rather than simply a

  

Source: Andrews, Peter B. - Department of Mathematical Sciences, Carnegie Mellon University

 

Collections: Mathematics