Novel Bounds on the Probability of Misclassification in Majority Voting: Leveraging the Majority Size

A.T.J.R. Cobbenhagen; A. Carè; M. C. Campi; F. A. Ramponi; D. J. Antunes; W.P.M.H. Heemels

doi:10.1109/LCSYS.2020.3040961

Novel Bounds on the Probability of Misclassification in Majority Voting: Leveraging the Majority Size

A.T.J.R. Cobbenhagen (Corresponding author), A. Carè, M. C. Campi, F. A. Ramponi, D. J. Antunes, W.P.M.H. Heemels

Research output: Contribution to journal › Article › Academic › peer-review

Abstract

Majority voting is often employed as a tool to increase the robustness of data-driven decisions and control policies, a fact which calls for rigorous, quantitative evaluations of the limits and the potentials of majority voting schemes. This letter focuses on the case where the voting agents are binary classifiers and introduces novel bounds on the probability of misclassification conditioned on the size of the majority. We show that these bounds can be much smaller than the traditional upper bounds on the probability of misclassification. These bounds can be used in a 'Probably Approximately Correct' (PAC) setting, which allows for a practical implementation.

Original language	English
Article number	9272617
Pages (from-to)	1513-1518
Number of pages	6
Journal	IEEE Control Systems Letters
Volume	5
Issue number	5
DOIs	https://doi.org/10.1109/LCSYS.2020.3040961
Publication status	Published - Nov 2021

Keywords

agents-based systems
Machine learning
statistical learning

Access to Document

10.1109/LCSYS.2020.3040961

Link to publication in Scopus

Fingerprint Dive into the research topics of 'Novel Bounds on the Probability of Misclassification in Majority Voting: Leveraging the Majority Size'. Together they form a unique fingerprint.

Cite this

@article{e8a263fa70da4c17879e02b56f9c7de8,

title = "Novel Bounds on the Probability of Misclassification in Majority Voting: Leveraging the Majority Size",

abstract = "Majority voting is often employed as a tool to increase the robustness of data-driven decisions and control policies, a fact which calls for rigorous, quantitative evaluations of the limits and the potentials of majority voting schemes. This letter focuses on the case where the voting agents are binary classifiers and introduces novel bounds on the probability of misclassification conditioned on the size of the majority. We show that these bounds can be much smaller than the traditional upper bounds on the probability of misclassification. These bounds can be used in a 'Probably Approximately Correct' (PAC) setting, which allows for a practical implementation. ",

keywords = "agents-based systems, Machine learning, statistical learning",

author = "A.T.J.R. Cobbenhagen and A. Car{\`e} and Campi, {M. C.} and Ramponi, {F. A.} and Antunes, {D. J.} and W.P.M.H. Heemels",

year = "2021",

month = nov,

doi = "10.1109/LCSYS.2020.3040961",

language = "English",

volume = "5",

pages = "1513--1518",

journal = "IEEE Control Systems Letters",

issn = "2475-1456",

publisher = "Institute of Electrical and Electronics Engineers",

number = "5",

}

Novel Bounds on the Probability of Misclassification in Majority Voting: Leveraging the Majority Size. / Cobbenhagen, A.T.J.R. (Corresponding author); Carè, A.; Campi, M. C.; Ramponi, F. A.; Antunes, D. J.; Heemels, W.P.M.H.

In: IEEE Control Systems Letters, Vol. 5, No. 5, 9272617, 11.2021, p. 1513-1518.

Research output: Contribution to journal › Article › Academic › peer-review

TY - JOUR

T1 - Novel Bounds on the Probability of Misclassification in Majority Voting: Leveraging the Majority Size

AU - Cobbenhagen, A.T.J.R.

AU - Carè, A.

AU - Campi, M. C.

AU - Ramponi, F. A.

AU - Antunes, D. J.

AU - Heemels, W.P.M.H.

PY - 2021/11

Y1 - 2021/11

N2 - Majority voting is often employed as a tool to increase the robustness of data-driven decisions and control policies, a fact which calls for rigorous, quantitative evaluations of the limits and the potentials of majority voting schemes. This letter focuses on the case where the voting agents are binary classifiers and introduces novel bounds on the probability of misclassification conditioned on the size of the majority. We show that these bounds can be much smaller than the traditional upper bounds on the probability of misclassification. These bounds can be used in a 'Probably Approximately Correct' (PAC) setting, which allows for a practical implementation.

AB - Majority voting is often employed as a tool to increase the robustness of data-driven decisions and control policies, a fact which calls for rigorous, quantitative evaluations of the limits and the potentials of majority voting schemes. This letter focuses on the case where the voting agents are binary classifiers and introduces novel bounds on the probability of misclassification conditioned on the size of the majority. We show that these bounds can be much smaller than the traditional upper bounds on the probability of misclassification. These bounds can be used in a 'Probably Approximately Correct' (PAC) setting, which allows for a practical implementation.

KW - agents-based systems

KW - Machine learning

KW - statistical learning

UR - http://www.scopus.com/inward/record.url?scp=85098552590&partnerID=8YFLogxK

U2 - 10.1109/LCSYS.2020.3040961

DO - 10.1109/LCSYS.2020.3040961

M3 - Article

AN - SCOPUS:85098552590

VL - 5

SP - 1513

EP - 1518

JO - IEEE Control Systems Letters

JF - IEEE Control Systems Letters

SN - 2475-1456

IS - 5

M1 - 9272617

ER -