Structural and parametric uncertainties in full Bayesian and graphical lasso based approaches

Beyond edge weights in psychological networks

Gabor Hullam, G. Juhász, Bill Deakin, Peter Antal

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Uncertainty over model structures poses a challenge for many approaches exploring effect strength parameters at system-level. Monte Carlo methods for full Bayesian model averaging over model structures require considerable computational resources, whereas bootstrapped graphical lasso and its approximations offer scalable alternatives with lower complexity. Although the computational efficiency of graphical lasso based approaches has prompted growing number of applications, the restrictive assumptions of this approach are frequently ignored. We demonstrate using an artificial and a real-world example that full Bayesian averaging using Bayesian networks provides detailed estimates through posterior distributions for structural and parametric uncertainties and it is a feasible alternative, which is routinely applicable in mid-sized biomedical problems with hundreds of variables. We compare Bayesian estimates with corresponding frequentist quantities from bootstrapped graphical lasso using pairwise Markov Random Fields, discussing also their different interpretations. We present results using synthetic data from an artificial model and using the UK Biobank data set to construct a psychopathological network centered around depression (this research has been conducted using the UK Biobank Resource under Application Number 1602).

Original languageEnglish
Title of host publication2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781467389884
DOIs
Publication statusPublished - Oct 4 2017
Event2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2017 - Manchester, United Kingdom
Duration: Aug 23 2017Aug 25 2017

Other

Other2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2017
CountryUnited Kingdom
CityManchester
Period8/23/178/25/17

Fingerprint

Lasso
Parametric Uncertainty
Model structures
Uncertainty
uncertainty
Psychology
Weights and Measures
Bayesian networks
Computational efficiency
Bayesian Model Averaging
Monte Carlo Method
Resources
Monte Carlo methods
Alternatives
Synthetic Data
Posterior distribution
Bayesian Networks
Computational Efficiency
Estimate
Monte Carlo method

ASJC Scopus subject areas

  • Computational Mathematics
  • Modelling and Simulation
  • Health Informatics
  • Agricultural and Biological Sciences (miscellaneous)
  • Artificial Intelligence
  • Computer Science Applications
  • Computer Vision and Pattern Recognition

Cite this

Hullam, G., Juhász, G., Deakin, B., & Antal, P. (2017). Structural and parametric uncertainties in full Bayesian and graphical lasso based approaches: Beyond edge weights in psychological networks. In 2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2017 [8058566] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CIBCB.2017.8058566

Structural and parametric uncertainties in full Bayesian and graphical lasso based approaches : Beyond edge weights in psychological networks. / Hullam, Gabor; Juhász, G.; Deakin, Bill; Antal, Peter.

2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2017. Institute of Electrical and Electronics Engineers Inc., 2017. 8058566.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Hullam, G, Juhász, G, Deakin, B & Antal, P 2017, Structural and parametric uncertainties in full Bayesian and graphical lasso based approaches: Beyond edge weights in psychological networks. in 2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2017., 8058566, Institute of Electrical and Electronics Engineers Inc., 2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2017, Manchester, United Kingdom, 8/23/17. https://doi.org/10.1109/CIBCB.2017.8058566
Hullam G, Juhász G, Deakin B, Antal P. Structural and parametric uncertainties in full Bayesian and graphical lasso based approaches: Beyond edge weights in psychological networks. In 2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2017. Institute of Electrical and Electronics Engineers Inc. 2017. 8058566 https://doi.org/10.1109/CIBCB.2017.8058566
Hullam, Gabor ; Juhász, G. ; Deakin, Bill ; Antal, Peter. / Structural and parametric uncertainties in full Bayesian and graphical lasso based approaches : Beyond edge weights in psychological networks. 2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2017. Institute of Electrical and Electronics Engineers Inc., 2017.
@inproceedings{45ecaa47c45f4128a2a6fc7efc040f1a,
title = "Structural and parametric uncertainties in full Bayesian and graphical lasso based approaches: Beyond edge weights in psychological networks",
abstract = "Uncertainty over model structures poses a challenge for many approaches exploring effect strength parameters at system-level. Monte Carlo methods for full Bayesian model averaging over model structures require considerable computational resources, whereas bootstrapped graphical lasso and its approximations offer scalable alternatives with lower complexity. Although the computational efficiency of graphical lasso based approaches has prompted growing number of applications, the restrictive assumptions of this approach are frequently ignored. We demonstrate using an artificial and a real-world example that full Bayesian averaging using Bayesian networks provides detailed estimates through posterior distributions for structural and parametric uncertainties and it is a feasible alternative, which is routinely applicable in mid-sized biomedical problems with hundreds of variables. We compare Bayesian estimates with corresponding frequentist quantities from bootstrapped graphical lasso using pairwise Markov Random Fields, discussing also their different interpretations. We present results using synthetic data from an artificial model and using the UK Biobank data set to construct a psychopathological network centered around depression (this research has been conducted using the UK Biobank Resource under Application Number 1602).",
author = "Gabor Hullam and G. Juh{\'a}sz and Bill Deakin and Peter Antal",
year = "2017",
month = "10",
day = "4",
doi = "10.1109/CIBCB.2017.8058566",
language = "English",
booktitle = "2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Structural and parametric uncertainties in full Bayesian and graphical lasso based approaches

T2 - Beyond edge weights in psychological networks

AU - Hullam, Gabor

AU - Juhász, G.

AU - Deakin, Bill

AU - Antal, Peter

PY - 2017/10/4

Y1 - 2017/10/4

N2 - Uncertainty over model structures poses a challenge for many approaches exploring effect strength parameters at system-level. Monte Carlo methods for full Bayesian model averaging over model structures require considerable computational resources, whereas bootstrapped graphical lasso and its approximations offer scalable alternatives with lower complexity. Although the computational efficiency of graphical lasso based approaches has prompted growing number of applications, the restrictive assumptions of this approach are frequently ignored. We demonstrate using an artificial and a real-world example that full Bayesian averaging using Bayesian networks provides detailed estimates through posterior distributions for structural and parametric uncertainties and it is a feasible alternative, which is routinely applicable in mid-sized biomedical problems with hundreds of variables. We compare Bayesian estimates with corresponding frequentist quantities from bootstrapped graphical lasso using pairwise Markov Random Fields, discussing also their different interpretations. We present results using synthetic data from an artificial model and using the UK Biobank data set to construct a psychopathological network centered around depression (this research has been conducted using the UK Biobank Resource under Application Number 1602).

AB - Uncertainty over model structures poses a challenge for many approaches exploring effect strength parameters at system-level. Monte Carlo methods for full Bayesian model averaging over model structures require considerable computational resources, whereas bootstrapped graphical lasso and its approximations offer scalable alternatives with lower complexity. Although the computational efficiency of graphical lasso based approaches has prompted growing number of applications, the restrictive assumptions of this approach are frequently ignored. We demonstrate using an artificial and a real-world example that full Bayesian averaging using Bayesian networks provides detailed estimates through posterior distributions for structural and parametric uncertainties and it is a feasible alternative, which is routinely applicable in mid-sized biomedical problems with hundreds of variables. We compare Bayesian estimates with corresponding frequentist quantities from bootstrapped graphical lasso using pairwise Markov Random Fields, discussing also their different interpretations. We present results using synthetic data from an artificial model and using the UK Biobank data set to construct a psychopathological network centered around depression (this research has been conducted using the UK Biobank Resource under Application Number 1602).

UR - http://www.scopus.com/inward/record.url?scp=85034652013&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85034652013&partnerID=8YFLogxK

U2 - 10.1109/CIBCB.2017.8058566

DO - 10.1109/CIBCB.2017.8058566

M3 - Conference contribution

BT - 2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2017

PB - Institute of Electrical and Electronics Engineers Inc.

ER -