A new method for constructing networks from binary data

Claudia D van Borkulo, Denny Borsboom, Sacha Epskamp, Tessa F Blanken, Lynn Boschloo, Robert A Schoevers, Lourens J Waldorp

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

Network analysis is entering fields where network structures are unknown, such as psychology and the educational sciences. A crucial step in the application of network models lies in the assessment of network structure. Current methods either have serious drawbacks or are only suitable for Gaussian data. In the present paper, we present a method for assessing network structures from binary data. Although models for binary data are infamous for their computational intractability, we present a computationally efficient model for estimating network structures. The approach, which is based on Ising models as used in physics, combines logistic regression with model selection based on a Goodness-of-Fit measure to identify relevant relationships between variables that define connections in a network. A validation study shows that this method succeeds in revealing the most relevant features of a network for realistic sample sizes. We apply our proposed method to estimate the network of depression and anxiety symptoms from symptom scores of 1108 subjects. Possible extensions of the model are discussed.

Original languageEnglish
Pages (from-to)5918
JournalScientific Reports
Volume4
DOIs
Publication statusPublished - 1 Aug 2014

Fingerprint

Educational Psychology
Validation Studies
Physics
Sample Size
Anxiety
Logistic Models
Depression

Keywords

  • Algorithms
  • Case-Control Studies
  • Computer Simulation
  • Depression/diagnosis
  • Humans
  • Models, Theoretical
  • Software

Cite this

van Borkulo, C. D., Borsboom, D., Epskamp, S., Blanken, T. F., Boschloo, L., Schoevers, R. A., & Waldorp, L. J. (2014). A new method for constructing networks from binary data. Scientific Reports, 4, 5918. https://doi.org/10.1038/srep05918
van Borkulo, Claudia D ; Borsboom, Denny ; Epskamp, Sacha ; Blanken, Tessa F ; Boschloo, Lynn ; Schoevers, Robert A ; Waldorp, Lourens J. / A new method for constructing networks from binary data. In: Scientific Reports. 2014 ; Vol. 4. pp. 5918.
@article{ed98196398184a7a87dc0279f50c11e4,
title = "A new method for constructing networks from binary data",
abstract = "Network analysis is entering fields where network structures are unknown, such as psychology and the educational sciences. A crucial step in the application of network models lies in the assessment of network structure. Current methods either have serious drawbacks or are only suitable for Gaussian data. In the present paper, we present a method for assessing network structures from binary data. Although models for binary data are infamous for their computational intractability, we present a computationally efficient model for estimating network structures. The approach, which is based on Ising models as used in physics, combines logistic regression with model selection based on a Goodness-of-Fit measure to identify relevant relationships between variables that define connections in a network. A validation study shows that this method succeeds in revealing the most relevant features of a network for realistic sample sizes. We apply our proposed method to estimate the network of depression and anxiety symptoms from symptom scores of 1108 subjects. Possible extensions of the model are discussed.",
keywords = "Algorithms, Case-Control Studies, Computer Simulation, Depression/diagnosis, Humans, Models, Theoretical, Software",
author = "{van Borkulo}, {Claudia D} and Denny Borsboom and Sacha Epskamp and Blanken, {Tessa F} and Lynn Boschloo and Schoevers, {Robert A} and Waldorp, {Lourens J}",
year = "2014",
month = "8",
day = "1",
doi = "10.1038/srep05918",
language = "English",
volume = "4",
pages = "5918",
journal = "Scientific Reports",
issn = "2045-2322",
publisher = "Nature Publishing Group",

}

van Borkulo, CD, Borsboom, D, Epskamp, S, Blanken, TF, Boschloo, L, Schoevers, RA & Waldorp, LJ 2014, 'A new method for constructing networks from binary data' Scientific Reports, vol. 4, pp. 5918. https://doi.org/10.1038/srep05918

A new method for constructing networks from binary data. / van Borkulo, Claudia D; Borsboom, Denny; Epskamp, Sacha; Blanken, Tessa F; Boschloo, Lynn; Schoevers, Robert A; Waldorp, Lourens J.

In: Scientific Reports, Vol. 4, 01.08.2014, p. 5918.

Research output: Contribution to JournalArticleAcademicpeer-review

TY - JOUR

T1 - A new method for constructing networks from binary data

AU - van Borkulo, Claudia D

AU - Borsboom, Denny

AU - Epskamp, Sacha

AU - Blanken, Tessa F

AU - Boschloo, Lynn

AU - Schoevers, Robert A

AU - Waldorp, Lourens J

PY - 2014/8/1

Y1 - 2014/8/1

N2 - Network analysis is entering fields where network structures are unknown, such as psychology and the educational sciences. A crucial step in the application of network models lies in the assessment of network structure. Current methods either have serious drawbacks or are only suitable for Gaussian data. In the present paper, we present a method for assessing network structures from binary data. Although models for binary data are infamous for their computational intractability, we present a computationally efficient model for estimating network structures. The approach, which is based on Ising models as used in physics, combines logistic regression with model selection based on a Goodness-of-Fit measure to identify relevant relationships between variables that define connections in a network. A validation study shows that this method succeeds in revealing the most relevant features of a network for realistic sample sizes. We apply our proposed method to estimate the network of depression and anxiety symptoms from symptom scores of 1108 subjects. Possible extensions of the model are discussed.

AB - Network analysis is entering fields where network structures are unknown, such as psychology and the educational sciences. A crucial step in the application of network models lies in the assessment of network structure. Current methods either have serious drawbacks or are only suitable for Gaussian data. In the present paper, we present a method for assessing network structures from binary data. Although models for binary data are infamous for their computational intractability, we present a computationally efficient model for estimating network structures. The approach, which is based on Ising models as used in physics, combines logistic regression with model selection based on a Goodness-of-Fit measure to identify relevant relationships between variables that define connections in a network. A validation study shows that this method succeeds in revealing the most relevant features of a network for realistic sample sizes. We apply our proposed method to estimate the network of depression and anxiety symptoms from symptom scores of 1108 subjects. Possible extensions of the model are discussed.

KW - Algorithms

KW - Case-Control Studies

KW - Computer Simulation

KW - Depression/diagnosis

KW - Humans

KW - Models, Theoretical

KW - Software

U2 - 10.1038/srep05918

DO - 10.1038/srep05918

M3 - Article

VL - 4

SP - 5918

JO - Scientific Reports

JF - Scientific Reports

SN - 2045-2322

ER -