A New Partially Segment-Wise Coupled Piece-Wise Linear Regression Model for Statistical Network Structure Inference

Mahdi Shafiee Kamalabad; Marco Grzegorczyk

doi:10.1007/978-3-030-34585-3_13

A New Partially Segment-Wise Coupled Piece-Wise Linear Regression Model for Statistical Network Structure Inference

Mahdi Shafiee Kamalabad, Marco Grzegorczyk^*

^*Corresponding author for this work

Probability and Statistics

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

93 Downloads (Pure)

Abstract

We propose a new non-homogeneous dynamic Bayesian network with partially segment-wise sequentially coupled network parameters. The idea is to infer the segmentation of a time series of network data using multiple changepoint processes, and to model the data in each segment by linear regression models. The conventional uncoupled models infer the network interaction parameters for each segment separately, without any systematic information-sharing among segments. More recently, it was proposed to couple the network interaction parameters sequentially among segments. The idea is to enforce the parameters of any segment to stay similar to those of the previous segment. This coupling mechanism can be disadvantageous, as it enforces coupling and does not feature any options to uncouple. We propose a new consensus model that infers for each individual segment whether it should be coupled to (or better should stay uncoupled from) the preceding one.

Original language	English
Title of host publication	Computational Intelligence Methods for Bioinformatics and Biostatistics
Subtitle of host publication	15th International Meeting, CIBB 2018, Caparica, Portugal, September 6–8, 2018, Revised Selected Papers
Editors	Maria Raposo, Paulo Ribeiro, Susana Sério, Antonino Staiano, Angelo Ciaramella
Publisher	Springer
Pages	139-152
Number of pages	14
ISBN (Electronic)	978-3-030-34585-3
ISBN (Print)	978-3-030-34584-6
DOIs	https://doi.org/10.1007/978-3-030-34585-3_13
Publication status	Published - 2020
Event	15th International Meeting, CIBB 2018 - Caparica, Portugal Duration: 6-Sept-2018 → 8-Sept-2018

Publication series

Name	Lecture Notes in Bioinformatics
Publisher	Springer
Number	1

Conference

Conference	15th International Meeting, CIBB 2018
Country/Territory	Portugal
City	Caparica
Period	06/09/2018 → 08/09/2018

Keywords

Bayesian piece-wise linear regression
Dynamic Bayesian networks
Network structure learning
Partial segment-wise coupling

Access to Document

10.1007/978-3-030-34585-3_13

A New Partially Segment-Wise Coupled Piece-Wise Linear Regression Model for Statistical Network Structure InferenceFinal publisher's version, 692 KBLicence: Taverne

Handle.net

Persistent link

Cite this

Shafiee Kamalabad, M., & Grzegorczyk, M. (2020). A New Partially Segment-Wise Coupled Piece-Wise Linear Regression Model for Statistical Network Structure Inference. In M. Raposo, P. Ribeiro, S. Sério, A. Staiano, & A. Ciaramella (Eds.), Computational Intelligence Methods for Bioinformatics and Biostatistics: 15th International Meeting, CIBB 2018, Caparica, Portugal, September 6–8, 2018, Revised Selected Papers (pp. 139-152). (Lecture Notes in Bioinformatics; No. 1). Springer. https://doi.org/10.1007/978-3-030-34585-3_13

Shafiee Kamalabad, Mahdi ; Grzegorczyk, Marco. / A New Partially Segment-Wise Coupled Piece-Wise Linear Regression Model for Statistical Network Structure Inference. Computational Intelligence Methods for Bioinformatics and Biostatistics: 15th International Meeting, CIBB 2018, Caparica, Portugal, September 6–8, 2018, Revised Selected Papers. editor / Maria Raposo ; Paulo Ribeiro ; Susana Sério ; Antonino Staiano ; Angelo Ciaramella. Springer, 2020. pp. 139-152 (Lecture Notes in Bioinformatics; 1).

@inproceedings{c5b6283e722f446597f2c71c7912c34f,

title = "A New Partially Segment-Wise Coupled Piece-Wise Linear Regression Model for Statistical Network Structure Inference",

abstract = "We propose a new non-homogeneous dynamic Bayesian network with partially segment-wise sequentially coupled network parameters. The idea is to infer the segmentation of a time series of network data using multiple changepoint processes, and to model the data in each segment by linear regression models. The conventional uncoupled models infer the network interaction parameters for each segment separately, without any systematic information-sharing among segments. More recently, it was proposed to couple the network interaction parameters sequentially among segments. The idea is to enforce the parameters of any segment to stay similar to those of the previous segment. This coupling mechanism can be disadvantageous, as it enforces coupling and does not feature any options to uncouple. We propose a new consensus model that infers for each individual segment whether it should be coupled to (or better should stay uncoupled from) the preceding one.",

keywords = "Bayesian piece-wise linear regression, Dynamic Bayesian networks, Network structure learning, Partial segment-wise coupling",

author = "{Shafiee Kamalabad}, Mahdi and Marco Grzegorczyk",

year = "2020",

doi = "10.1007/978-3-030-34585-3_13",

language = "English",

isbn = "978-3-030-34584-6",

series = "Lecture Notes in Bioinformatics",

publisher = "Springer",

number = "1",

pages = "139--152",

editor = "Maria Raposo and Paulo Ribeiro and Susana S{\'e}rio and Antonino Staiano and Angelo Ciaramella",

booktitle = "Computational Intelligence Methods for Bioinformatics and Biostatistics",

note = "15th International Meeting, CIBB 2018 ; Conference date: 06-09-2018 Through 08-09-2018",

}

Shafiee Kamalabad, M & Grzegorczyk, M 2020, A New Partially Segment-Wise Coupled Piece-Wise Linear Regression Model for Statistical Network Structure Inference. in M Raposo, P Ribeiro, S Sério, A Staiano & A Ciaramella (eds), Computational Intelligence Methods for Bioinformatics and Biostatistics: 15th International Meeting, CIBB 2018, Caparica, Portugal, September 6–8, 2018, Revised Selected Papers. Lecture Notes in Bioinformatics, no. 1, Springer, pp. 139-152, 15th International Meeting, CIBB 2018, Caparica, Portugal, 06/09/2018. https://doi.org/10.1007/978-3-030-34585-3_13

A New Partially Segment-Wise Coupled Piece-Wise Linear Regression Model for Statistical Network Structure Inference. / Shafiee Kamalabad, Mahdi; Grzegorczyk, Marco.
Computational Intelligence Methods for Bioinformatics and Biostatistics: 15th International Meeting, CIBB 2018, Caparica, Portugal, September 6–8, 2018, Revised Selected Papers. ed. / Maria Raposo; Paulo Ribeiro; Susana Sério; Antonino Staiano; Angelo Ciaramella. Springer, 2020. p. 139-152 (Lecture Notes in Bioinformatics; No. 1).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - A New Partially Segment-Wise Coupled Piece-Wise Linear Regression Model for Statistical Network Structure Inference

AU - Shafiee Kamalabad, Mahdi

AU - Grzegorczyk, Marco

PY - 2020

Y1 - 2020

N2 - We propose a new non-homogeneous dynamic Bayesian network with partially segment-wise sequentially coupled network parameters. The idea is to infer the segmentation of a time series of network data using multiple changepoint processes, and to model the data in each segment by linear regression models. The conventional uncoupled models infer the network interaction parameters for each segment separately, without any systematic information-sharing among segments. More recently, it was proposed to couple the network interaction parameters sequentially among segments. The idea is to enforce the parameters of any segment to stay similar to those of the previous segment. This coupling mechanism can be disadvantageous, as it enforces coupling and does not feature any options to uncouple. We propose a new consensus model that infers for each individual segment whether it should be coupled to (or better should stay uncoupled from) the preceding one.

AB - We propose a new non-homogeneous dynamic Bayesian network with partially segment-wise sequentially coupled network parameters. The idea is to infer the segmentation of a time series of network data using multiple changepoint processes, and to model the data in each segment by linear regression models. The conventional uncoupled models infer the network interaction parameters for each segment separately, without any systematic information-sharing among segments. More recently, it was proposed to couple the network interaction parameters sequentially among segments. The idea is to enforce the parameters of any segment to stay similar to those of the previous segment. This coupling mechanism can be disadvantageous, as it enforces coupling and does not feature any options to uncouple. We propose a new consensus model that infers for each individual segment whether it should be coupled to (or better should stay uncoupled from) the preceding one.

KW - Bayesian piece-wise linear regression

KW - Dynamic Bayesian networks

KW - Network structure learning

KW - Partial segment-wise coupling

UR - https://www.springer.com/de/book/9783030345846

U2 - 10.1007/978-3-030-34585-3_13

DO - 10.1007/978-3-030-34585-3_13

M3 - Conference contribution

SN - 978-3-030-34584-6

T3 - Lecture Notes in Bioinformatics

SP - 139

EP - 152

BT - Computational Intelligence Methods for Bioinformatics and Biostatistics

A2 - Raposo, Maria

A2 - Ribeiro, Paulo

A2 - Sério, Susana

A2 - Staiano, Antonino

A2 - Ciaramella, Angelo

PB - Springer

T2 - 15th International Meeting, CIBB 2018

Y2 - 6 September 2018 through 8 September 2018

ER -

Shafiee Kamalabad M, Grzegorczyk M. A New Partially Segment-Wise Coupled Piece-Wise Linear Regression Model for Statistical Network Structure Inference. In Raposo M, Ribeiro P, Sério S, Staiano A, Ciaramella A, editors, Computational Intelligence Methods for Bioinformatics and Biostatistics: 15th International Meeting, CIBB 2018, Caparica, Portugal, September 6–8, 2018, Revised Selected Papers. Springer. 2020. p. 139-152. (Lecture Notes in Bioinformatics; 1). doi: 10.1007/978-3-030-34585-3_13

A New Partially Segment-Wise Coupled Piece-Wise Linear Regression Model for Statistical Network Structure Inference

Abstract

Publication series

Conference

Keywords

Access to Document

Handle.net

Other files and links

Fingerprint

Cite this