QUT ePrints

Expert Elicitation for Bayesian Classification Trees

O'Leary, Rebecca A., Murray, Justine V., Low Choy, Samantha J., & Mengersen, Kerrie L. (2008) Expert Elicitation for Bayesian Classification Trees. Journal of Applied Probability & Statistics, 3(1), pp. 95-106.

[img] PDF (185kB)
Administrators only

    Abstract

    An expert elicitation approach for Bayesian classification trees is developed in this paper. This approach is illustrated for habitat suitability modelling of the threatened Australian brush-tailed rock-wallaby Petrogale penicillata, in which the opinion of one expert is elicited. In the ecological field, expert opinion has been acknowledged as providing valuable information in modelling, particularly when the observed data are limited or unreliable. The elicitation questions are on the size of the tree representing the number of decisions; the relative importance of the variables; and the splitting rules for the most important variables which quantify how decisions relate to variables. For each of these questions three approaches to elicitation were used: order from most important to least important; grade each item from 1 to 5; attach a numeric weight to each item. The results indicate that for this case study, expert informed priors were able to influence the tree structure (tree size, variables and splitting rules). Furthermore, through applying these priors a tree with better predictions of the presences was identified, compared to that based on non-informative priors. Hence combining expert informed priors with observed data using Bayesian classification trees may improve scientific understanding and conservation management planning.

    Impact and interest:

    Citation countsare sourced monthly from Scopus and Web of Science® citation databases.

    These databases contain citations from different subsets of available publications and different time periods and thus the citation count from each is usually different. Some works are not in either database and no count is displayed. Scopus includes citations from articles published in 1996 onwards, and Web of Science® generally from 1980 onwards.

    Citations counts from the Google Scholar™ indexing service can be viewed at the linked Google Scholar™ search.

    ID Code: 15671
    Item Type: Journal Article
    Additional Information: Access to the author-version is currently restricted pending permission from the publisher. For more information, please refer to the journal's website (see hypertext link) or contact the author.
    Additional URLs:
    Keywords: Expert elicitation, informative priors, Bayesian classification trees, Bayesian model, habitat suitability modelling, threatened species
    ISSN: 1930-6792
    Subjects: Australian and New Zealand Standard Research Classification > ENVIRONMENTAL SCIENCES (050000) > ENVIRONMENTAL SCIENCE AND MANAGEMENT (050200) > Conservation and Biodiversity (050202)
    Australian and New Zealand Standard Research Classification > PSYCHOLOGY AND COGNITIVE SCIENCES (170000) > COGNITIVE SCIENCE (170200) > Knowledge Representation and Machine Learning (170203)
    Australian and New Zealand Standard Research Classification > ENVIRONMENTAL SCIENCES (050000) > ECOLOGICAL APPLICATIONS (050100) > Landscape Ecology (050104)
    Australian and New Zealand Standard Research Classification > MATHEMATICAL SCIENCES (010000) > STATISTICS (010400) > Applied Statistics (010401)
    Divisions: Past > QUT Faculties & Divisions > Faculty of Science and Technology
    Copyright Owner: Copyright 2008 Dixie W Publishing Corporation
    Deposited On: 20 Nov 2008
    Last Modified: 03 Sep 2012 09:00

    Export: EndNote | Dublin Core | BibTeX

    Repository Staff Only: item control page