BSU Seminar: "Graphical and summary diagnostics for node level adequacy in Bayesian hierarchical models"

Add to your list(s) Download to your calendar using vCal

• Ida Scheel, University of Oslo
• Tuesday 22 October 2024, 14:00-15:00
• MRC Biostatistics Unit, East Forvie Building, Forvie Site Robinson Way Cambridge CB2 0SR..

If you have a question about this talk, please contact Alison Quenault.

This will be a free hybrid seminar. To register to attend virtually, please click here: https://cam-ac-uk.zoom.us/meeting/register/tZwkf-ysqDgqG9W5MbRoGxq5re7R9XadbBBd

For Bayesian hierarchical models represented by directed acyclic graphs, all neighbours of a node provide, potentially conflicting, information on the node. Hence identifying potential conflict between these different information sources can be important for assessing the adequacy of Bayesian hierarchical models. Building on this idea originating from [1], [3] constructed node based conflict measures that have been shown to be well calibrated. Sharing the idea of considering potential conflict between separate information sources for each node, [4] constructed a graphical diagnostic lcp. This can be used both to identify conflict at each node, and provide insight into the nature of the conflict, hence being more informative than summary diagnostics. We link these two ideas together by constructing a new diagnostic plot iic-lcp that is supplementary to lcp, but builds on the framework of [3]. It has the advantage over the lcp with the possibility to display curves corresponding to different parameters in the same plot, saving space and easing comparisons, particularly useful for sets of parameters representing exchangeable groups. We show how to visually read some of the conflict measures defined in [3] from iic-lcp, combining graphical and summary diagnostics.

For Bayesian hierarchical models represented by directed acyclic graphs, all neighbours of a node provide, potentially conflicting, information on the node. Hence identifying potential conflict between these different information sources can be important for assessing the adequacy of Bayesian hierarchical models. Building on this idea originating from [1], [3] constructed node based conflict measures that have been shown to be well calibrated. Sharing the idea of considering potential conflict between separate information sources for each node, [4] constructed a graphical diagnostic lcp. This can be used both to identify conflict at each node, and provide insight into the nature of the conflict, hence being more informative than summary diagnostics. We link these two ideas together by constructing a new diagnostic plot iic-lcp that is supplementary to lcp, but builds on the framework of [3]. It has the advantage over the lcp with the possibility to display curves corresponding to different parameters in the same plot, saving space and easing comparisons, particularly useful for sets of parameters representing exchangeable groups. We show how to visually read some of the conflict measures defined in [3] from iic-lcp, combining graphical and summary diagnostics.

[1] O’Hagan, A. (2003). HSSS model criticism. In Green, P. J., Richardson, S., and Hjort, N. L. (eds.), Highly Structured Stochastic Systems, 423–444. Oxford: Oxford University Press. [2] Dahl, F. A., G Gasemyr, J., and Natvig, B. (2007). A robust conflict measure of inconsistencies in Bayesian hierarchical models. Scand. J. Stat., 34: 816–828. [3] Gasemyr, J. and Natvig, B. (2009). Extensions of a conflict measure of inconsistencies in Bayesian hierarchical models. Scand. J. Stat., 36: 822–838. [4] Scheel, I., Green, P. J., and Rougier, J. C. (2011). A graphical diagnostic for identifying influential model Choices in Bayesian hierarchical models. Scand. J. Stat., 38: 529–550.

This talk is part of the MRC Biostatistics Unit Seminars series.

This talk is included in these lists:

• All Talks (aka the CURE list)
• BHRU Annual Lecture 2015
• BHRU Annual Lecture 2016
• Biology
• Cambridge Centre for Data-Driven Discovery (C2D3)
• Cambridge Immunology
• Cambridge talks
• Cardiovascular Epidemiology Unit Special Seminars
• Chris Davis' list
• Department of Public Health and Primary Care
• Graduate-Seminars
• Interested Talks
• Life Sciences
• Life Sciences
• ME Seminar
• MRC Biostatistics Unit Seminars
• MRC Biostatistics Unit, East Forvie Building, Forvie Site Robinson Way Cambridge CB2 0SR.
• Primary Care
• PublicHealth@Cambridge
• Trust & Technology Initiative - interesting events
• bld31
• my_list
• ndk22's list
• ob366-ai4er
• other talks
• primary care
• rp587

Note that ex-directory lists are not shown.