Publication:
Informed Search for Learning Causal Structure

Simple item page
dc.contributor.advisorDavid Jensen
dc.contributor.authorTaylor, Brian J
dc.contributor.departmentUniversity of Massachusetts Amherst
dc.date2023-09-23T13:12:19.000
dc.date.accessioned2024-04-26T16:10:35Z
dc.date.available2024-04-26T16:10:35Z
dc.date.submittedSeptember
dc.date.submitted2015
dc.description.abstractOver the past twenty-five years, a large number of algorithms have been developed to learn the structure of causal graphical models. Many of these algorithms learn causal structures by analyzing the implications of observed conditional independence among variables that describe characteristics of the domain being analyzed. They do so by applying inference rules, data analysis operations such as the conditional independence tests, each of which can eliminate large parts of the space of possible causal structures. Results show that the sequence of inference rules used by PC, a widely applied algorithm for constraint-based learning of causal models, is effective but not optimal. This is because algorithms such as PC ignore the probability of the outcomes of these inference rules. We demonstrate how an alternative algorithm can reliably outperform PC by taking into account the probability of inference rule outcomes. Specifically we show that an informed search that bases the order of causal inference on a prior probability distribution over the space of causal constraints can generate a flexible sequence of analysis that efficiently identifies the same results as PC. This class of algorithms is able to outperform PC even under uniform or erroneous priors.
dc.description.degreeDoctor of Philosophy (PhD)
dc.description.departmentComputer Science
dc.identifier.doihttps://doi.org/10.7275/7538380.0
dc.identifier.orcidN/A
dc.identifier.urihttps://hdl.handle.net/20.500.14394/19765
dc.relation.urlhttps://scholarworks.umass.edu/cgi/viewcontent.cgi?article=1553&context=dissertations_2&unstamped=1
dc.source.statuspublished
dc.subjectCausality
dc.subjectCausal Learning
dc.subjectCausal Graphical Models
dc.subjectPC Algorithm
dc.subjectCausal Inference
dc.subjectArtificial Intelligence and Robotics
dc.titleInformed Search for Learning Causal Structure
dc.typeopenaccess
dc.typearticle
dc.typedissertation
digcom.contributor.authorisAuthorOfPublication|email:btaylor@cs.umass.edu|institution:University of Massachusetts Amherst|Taylor, Brian J
digcom.identifierdissertations_2/540
digcom.identifier.contextkey7538380
digcom.identifier.submissionpathdissertations_2/540
dspace.entity.typePublication
Files
Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
Btaylor_Dissertation.Informed_Search_For_Learning_Causal_Structure.v1.pdf
Size:
8.01 MB
Format:
Adobe Portable Document Format
Name: Btaylor_Dissertation.Informed_Search_For_Learning_Causal_Structure.v1.pdf
Size: 8.01 MB
Kind: Adobe PDF
Collections
Dissertations