Shapley Value Based Variable Interaction Networks for Data Stream Analysis.

Jan Zenisek; Sebastian Dorl; Dominik Falkner; Lukas Gaisberger; Stephan M. Winkler; Michael Affenzeller

doi:10.1007/978-3-031-25312-6_15

Shapley Value Based Variable Interaction Networks for Data Stream Analysis.

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

1 Citation (Scopus)

Abstract

Due to the growing use of machine learning models in many critical domains, ambitions to make the models and their predictions explainable have increased recently significantly as new research interest. In this paper, we present an extension to the machine learning based data mining technique of variable interaction networks, to improve their structural stability, which enables more meaningful analysis. To verify the feasibility of our approach and it’s capability to provide human-interpretable insights, we discuss the results of experiments with a set of challenging benchmark instances, as well as with real-world data from energy network monitoring.

Original language	English
Title of host publication	Computer Aided Systems Theory – EUROCAST 2022 - 18th International Conference, Revised Selected Papers
Editors	Roberto Moreno-Díaz, Franz Pichler, Alexis Quesada-Arencibia
Pages	130-138
Number of pages	9
DOIs	https://doi.org/10.1007/978-3-031-25312-6_15
Publication status	Published - 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13789 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 7 Affordable and Clean Energy

Keywords

Data stream analysis
Energy network resilience
Interpretable machine learning
Photovoltaic systems
Shapley value

Access to Document

10.1007/978-3-031-25312-6_15

Cite this

Zenisek, J., Dorl, S., Falkner, D., Gaisberger, L., Winkler, S. M., & Affenzeller, M. (2022). Shapley Value Based Variable Interaction Networks for Data Stream Analysis. In R. Moreno-Díaz, F. Pichler, & A. Quesada-Arencibia (Eds.), Computer Aided Systems Theory – EUROCAST 2022 - 18th International Conference, Revised Selected Papers (pp. 130-138). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13789 LNCS). https://doi.org/10.1007/978-3-031-25312-6_15

Zenisek, Jan ; Dorl, Sebastian ; Falkner, Dominik et al. / Shapley Value Based Variable Interaction Networks for Data Stream Analysis. Computer Aided Systems Theory – EUROCAST 2022 - 18th International Conference, Revised Selected Papers. editor / Roberto Moreno-Díaz ; Franz Pichler ; Alexis Quesada-Arencibia. 2022. pp. 130-138 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{1419afc902d74956be7ffb2fe2dde35d,

title = "Shapley Value Based Variable Interaction Networks for Data Stream Analysis.",

abstract = "Due to the growing use of machine learning models in many critical domains, ambitions to make the models and their predictions explainable have increased recently significantly as new research interest. In this paper, we present an extension to the machine learning based data mining technique of variable interaction networks, to improve their structural stability, which enables more meaningful analysis. To verify the feasibility of our approach and it{\textquoteright}s capability to provide human-interpretable insights, we discuss the results of experiments with a set of challenging benchmark instances, as well as with real-world data from energy network monitoring.",

keywords = "Data stream analysis, Energy network resilience, Interpretable machine learning, Photovoltaic systems, Shapley value",

author = "Jan Zenisek and Sebastian Dorl and Dominik Falkner and Lukas Gaisberger and Winkler, \{Stephan M.\} and Michael Affenzeller",

note = "Funding Information: Acknowledgments. The work described in this paper was done within the projects “RESINET”, funded by the European Fund for Regional Development (EFRE) and the country of Upper Austria as part of the program “Investing in Growth and Jobs 2014–2020” and “Secure Prescriptive Analytics”, funded by the country of Upper of Austria as part of the program “\#upperVISION2030”. Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.",

year = "2022",

doi = "10.1007/978-3-031-25312-6\_15",

language = "English",

isbn = "9783031253119",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

pages = "130--138",

editor = "Roberto Moreno-D{\'i}az and Franz Pichler and Alexis Quesada-Arencibia",

booktitle = "Computer Aided Systems Theory – EUROCAST 2022 - 18th International Conference, Revised Selected Papers",

}

Zenisek, J , Dorl, S, Falkner, D, Gaisberger, L , Winkler, SM & Affenzeller, M 2022, Shapley Value Based Variable Interaction Networks for Data Stream Analysis. in R Moreno-Díaz, F Pichler & A Quesada-Arencibia (eds), Computer Aided Systems Theory – EUROCAST 2022 - 18th International Conference, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13789 LNCS, pp. 130-138. https://doi.org/10.1007/978-3-031-25312-6_15

Shapley Value Based Variable Interaction Networks for Data Stream Analysis. / Zenisek, Jan ; Dorl, Sebastian; Falkner, Dominik et al.
Computer Aided Systems Theory – EUROCAST 2022 - 18th International Conference, Revised Selected Papers. ed. / Roberto Moreno-Díaz; Franz Pichler; Alexis Quesada-Arencibia. 2022. p. 130-138 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13789 LNCS).

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

TY - GEN

T1 - Shapley Value Based Variable Interaction Networks for Data Stream Analysis.

AU - Zenisek, Jan

AU - Dorl, Sebastian

AU - Falkner, Dominik

AU - Gaisberger, Lukas

AU - Winkler, Stephan M.

AU - Affenzeller, Michael

N1 - Funding Information: Acknowledgments. The work described in this paper was done within the projects “RESINET”, funded by the European Fund for Regional Development (EFRE) and the country of Upper Austria as part of the program “Investing in Growth and Jobs 2014–2020” and “Secure Prescriptive Analytics”, funded by the country of Upper of Austria as part of the program “#upperVISION2030”. Publisher Copyright: © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

PY - 2022

Y1 - 2022

N2 - Due to the growing use of machine learning models in many critical domains, ambitions to make the models and their predictions explainable have increased recently significantly as new research interest. In this paper, we present an extension to the machine learning based data mining technique of variable interaction networks, to improve their structural stability, which enables more meaningful analysis. To verify the feasibility of our approach and it’s capability to provide human-interpretable insights, we discuss the results of experiments with a set of challenging benchmark instances, as well as with real-world data from energy network monitoring.

AB - Due to the growing use of machine learning models in many critical domains, ambitions to make the models and their predictions explainable have increased recently significantly as new research interest. In this paper, we present an extension to the machine learning based data mining technique of variable interaction networks, to improve their structural stability, which enables more meaningful analysis. To verify the feasibility of our approach and it’s capability to provide human-interpretable insights, we discuss the results of experiments with a set of challenging benchmark instances, as well as with real-world data from energy network monitoring.

KW - Data stream analysis

KW - Energy network resilience

KW - Interpretable machine learning

KW - Photovoltaic systems

KW - Shapley value

UR - https://www.scopus.com/pages/publications/85151123862

U2 - 10.1007/978-3-031-25312-6_15

DO - 10.1007/978-3-031-25312-6_15

M3 - Conference contribution

SN - 9783031253119

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 130

EP - 138

BT - Computer Aided Systems Theory – EUROCAST 2022 - 18th International Conference, Revised Selected Papers

A2 - Moreno-Díaz, Roberto

A2 - Pichler, Franz

A2 - Quesada-Arencibia, Alexis

ER -

Zenisek J , Dorl S, Falkner D, Gaisberger L , Winkler SM , Affenzeller M. Shapley Value Based Variable Interaction Networks for Data Stream Analysis. In Moreno-Díaz R, Pichler F, Quesada-Arencibia A, editors, Computer Aided Systems Theory – EUROCAST 2022 - 18th International Conference, Revised Selected Papers. 2022. p. 130-138. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-25312-6_15

Shapley Value Based Variable Interaction Networks for Data Stream Analysis.

Abstract

Publication series

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this