Blind Multiclass Ensemble Classification

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

The rising interest in pattern recognition and data analytics has spurred the development of innovative machine learning algorithms and tools. However, as each algorithm has its strengths and limitations, one is motivated to judiciously fuse multiple algorithms in order to find the 'best' performing one, for a given dataset. Ensemble learning aims at such high-performance meta-algorithm, by combining the outputs from multiple algorithms. The present work introduces a blind scheme for learning from ensembles of classifiers, using a moment matching method that leverages joint tensor and matrix factorization. Blind refers to the combiner who has no knowledge of the ground-truth labels that each classifier has been trained on. A rigorous performance analysis is derived and the proposed scheme is evaluated on synthetic and real datasets.

Original languageEnglish (US)
Article number8421667
Pages (from-to)4737-4752
Number of pages16
JournalIEEE Transactions on Signal Processing
Volume66
Issue number18
DOIs
StatePublished - Sep 15 2018

Fingerprint

Classifiers
Electric fuses
Factorization
Learning algorithms
Pattern recognition
Tensors
Learning systems
Labels

Keywords

  • Ensemble learning
  • crowdsourcing
  • multiclass classification
  • unsupervised

Cite this

Blind Multiclass Ensemble Classification. / Traganitis, Panagiotis A.; Pages-Zamora, Alba; Giannakis, Georgios B.

In: IEEE Transactions on Signal Processing, Vol. 66, No. 18, 8421667, 15.09.2018, p. 4737-4752.

Research output: Contribution to journalArticle

@article{4405ef6d8bc342078c5ce099a7ef04cc,
title = "Blind Multiclass Ensemble Classification",
abstract = "The rising interest in pattern recognition and data analytics has spurred the development of innovative machine learning algorithms and tools. However, as each algorithm has its strengths and limitations, one is motivated to judiciously fuse multiple algorithms in order to find the 'best' performing one, for a given dataset. Ensemble learning aims at such high-performance meta-algorithm, by combining the outputs from multiple algorithms. The present work introduces a blind scheme for learning from ensembles of classifiers, using a moment matching method that leverages joint tensor and matrix factorization. Blind refers to the combiner who has no knowledge of the ground-truth labels that each classifier has been trained on. A rigorous performance analysis is derived and the proposed scheme is evaluated on synthetic and real datasets.",
keywords = "Ensemble learning, crowdsourcing, multiclass classification, unsupervised",
author = "Traganitis, {Panagiotis A.} and Alba Pages-Zamora and Giannakis, {Georgios B.}",
year = "2018",
month = "9",
day = "15",
doi = "10.1109/TSP.2018.2860562",
language = "English (US)",
volume = "66",
pages = "4737--4752",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "18",

}

TY - JOUR

T1 - Blind Multiclass Ensemble Classification

AU - Traganitis, Panagiotis A.

AU - Pages-Zamora, Alba

AU - Giannakis, Georgios B.

PY - 2018/9/15

Y1 - 2018/9/15

N2 - The rising interest in pattern recognition and data analytics has spurred the development of innovative machine learning algorithms and tools. However, as each algorithm has its strengths and limitations, one is motivated to judiciously fuse multiple algorithms in order to find the 'best' performing one, for a given dataset. Ensemble learning aims at such high-performance meta-algorithm, by combining the outputs from multiple algorithms. The present work introduces a blind scheme for learning from ensembles of classifiers, using a moment matching method that leverages joint tensor and matrix factorization. Blind refers to the combiner who has no knowledge of the ground-truth labels that each classifier has been trained on. A rigorous performance analysis is derived and the proposed scheme is evaluated on synthetic and real datasets.

AB - The rising interest in pattern recognition and data analytics has spurred the development of innovative machine learning algorithms and tools. However, as each algorithm has its strengths and limitations, one is motivated to judiciously fuse multiple algorithms in order to find the 'best' performing one, for a given dataset. Ensemble learning aims at such high-performance meta-algorithm, by combining the outputs from multiple algorithms. The present work introduces a blind scheme for learning from ensembles of classifiers, using a moment matching method that leverages joint tensor and matrix factorization. Blind refers to the combiner who has no knowledge of the ground-truth labels that each classifier has been trained on. A rigorous performance analysis is derived and the proposed scheme is evaluated on synthetic and real datasets.

KW - Ensemble learning

KW - crowdsourcing

KW - multiclass classification

KW - unsupervised

UR - http://www.scopus.com/inward/record.url?scp=85050730716&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85050730716&partnerID=8YFLogxK

U2 - 10.1109/TSP.2018.2860562

DO - 10.1109/TSP.2018.2860562

M3 - Article

AN - SCOPUS:85050730716

VL - 66

SP - 4737

EP - 4752

JO - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

IS - 18

M1 - 8421667

ER -