A new ensemble-based cascaded framework for multiclass training with simple weak learners

Teo Susnjak; Andre Barczak; Napoleon Reyes; Ken Hawick

doi:10.1007/978-3-642-23672-3_68

A new ensemble-based cascaded framework for multiclass training with simple weak learners

Teo Susnjak^*, Andre Barczak, Napoleon Reyes, Ken Hawick

^*Corresponding author for this work

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

2 Citations (Scopus)

Abstract

We present a novel approach to multiclass learning using an ensemble-based cascaded learning framework. By implementing a multiclass cascaded classifier with AdaBoost, we show how detection runtimes are accelerated since only a subset of the ensemble is executed, thus making the classifiers suitable for computer vision applications. We also propose a new multiclass weak learner and demonstrate the framework's ability to achieve arbitrarily low training errors in conjunction with it. We tested our algorithm against AdaBoost.OC, ECC and M2 multiclass learning methods, on seven benchmark UCI datasets. In our experiments, we found that our framework achieves higher accuracy on five out of seven datasets and displays faster runtime efficiency in all cases.

Original language	English
Title of host publication	Computer Analysis of Images and Patterns - 14th International Conference, CAIP 2011, Proceedings
Editors	Pedro Real, Daniel Diaz-Pernil, Helena Molina-Abril, Ainhoa Berciano, Walter Kropatsch
Publisher	Springer
Pages	563-570
Number of pages	8
Volume	PART 1
ISBN (Print)	9783642236716
DOIs	https://doi.org/10.1007/978-3-642-23672-3_68
Publication status	Published - 2011
Externally published	Yes
Event	14th International Conference on Computer Analysis of Images and Patterns, CAIP 2011 - Seville, Spain Duration: 29 Aug 2011 → 31 Aug 2011 Conference number: 14th

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 1
Volume	6854 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	14th International Conference on Computer Analysis of Images and Patterns, CAIP 2011
Abbreviated title	CAIP 2011
Country/Territory	Spain
City	Seville
Period	29/08/11 → 31/08/11

Access to Document

10.1007/978-3-642-23672-3_68

Cite this

Susnjak, T., Barczak, A., Reyes, N., & Hawick, K. (2011). A new ensemble-based cascaded framework for multiclass training with simple weak learners. In P. Real, D. Diaz-Pernil, H. Molina-Abril, A. Berciano, & W. Kropatsch (Eds.), Computer Analysis of Images and Patterns - 14th International Conference, CAIP 2011, Proceedings (Vol. PART 1, pp. 563-570). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6854 LNCS, No. PART 1). Springer. https://doi.org/10.1007/978-3-642-23672-3_68

Susnjak, Teo ; Barczak, Andre ; Reyes, Napoleon et al. / A new ensemble-based cascaded framework for multiclass training with simple weak learners. Computer Analysis of Images and Patterns - 14th International Conference, CAIP 2011, Proceedings. editor / Pedro Real ; Daniel Diaz-Pernil ; Helena Molina-Abril ; Ainhoa Berciano ; Walter Kropatsch. Vol. PART 1 Springer, 2011. pp. 563-570 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).

@inproceedings{4838456105b949378fe423bdc1c2df81,

title = "A new ensemble-based cascaded framework for multiclass training with simple weak learners",

abstract = "We present a novel approach to multiclass learning using an ensemble-based cascaded learning framework. By implementing a multiclass cascaded classifier with AdaBoost, we show how detection runtimes are accelerated since only a subset of the ensemble is executed, thus making the classifiers suitable for computer vision applications. We also propose a new multiclass weak learner and demonstrate the framework's ability to achieve arbitrarily low training errors in conjunction with it. We tested our algorithm against AdaBoost.OC, ECC and M2 multiclass learning methods, on seven benchmark UCI datasets. In our experiments, we found that our framework achieves higher accuracy on five out of seven datasets and displays faster runtime efficiency in all cases.",

author = "Teo Susnjak and Andre Barczak and Napoleon Reyes and Ken Hawick",

year = "2011",

doi = "10.1007/978-3-642-23672-3_68",

language = "English",

isbn = "9783642236716",

volume = "PART 1",

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

publisher = "Springer",

number = "PART 1",

pages = "563--570",

editor = "Pedro Real and Daniel Diaz-Pernil and Helena Molina-Abril and Ainhoa Berciano and Walter Kropatsch",

booktitle = "Computer Analysis of Images and Patterns - 14th International Conference, CAIP 2011, Proceedings",

address = "Germany",

note = "14th International Conference on Computer Analysis of Images and Patterns, CAIP 2011, CAIP 2011 ; Conference date: 29-08-2011 Through 31-08-2011",

}

Susnjak, T, Barczak, A, Reyes, N & Hawick, K 2011, A new ensemble-based cascaded framework for multiclass training with simple weak learners. in P Real, D Diaz-Pernil, H Molina-Abril, A Berciano & W Kropatsch (eds), Computer Analysis of Images and Patterns - 14th International Conference, CAIP 2011, Proceedings. vol. PART 1, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 6854 LNCS, Springer, pp. 563-570, 14th International Conference on Computer Analysis of Images and Patterns, CAIP 2011, Seville, Spain, 29/08/11. https://doi.org/10.1007/978-3-642-23672-3_68

A new ensemble-based cascaded framework for multiclass training with simple weak learners. / Susnjak, Teo; Barczak, Andre; Reyes, Napoleon et al.
Computer Analysis of Images and Patterns - 14th International Conference, CAIP 2011, Proceedings. ed. / Pedro Real; Daniel Diaz-Pernil; Helena Molina-Abril; Ainhoa Berciano; Walter Kropatsch. Vol. PART 1 Springer, 2011. p. 563-570 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6854 LNCS, No. PART 1).

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

TY - GEN

T1 - A new ensemble-based cascaded framework for multiclass training with simple weak learners

AU - Susnjak, Teo

AU - Barczak, Andre

AU - Reyes, Napoleon

AU - Hawick, Ken

N1 - Conference code: 14th

PY - 2011

Y1 - 2011

N2 - We present a novel approach to multiclass learning using an ensemble-based cascaded learning framework. By implementing a multiclass cascaded classifier with AdaBoost, we show how detection runtimes are accelerated since only a subset of the ensemble is executed, thus making the classifiers suitable for computer vision applications. We also propose a new multiclass weak learner and demonstrate the framework's ability to achieve arbitrarily low training errors in conjunction with it. We tested our algorithm against AdaBoost.OC, ECC and M2 multiclass learning methods, on seven benchmark UCI datasets. In our experiments, we found that our framework achieves higher accuracy on five out of seven datasets and displays faster runtime efficiency in all cases.

AB - We present a novel approach to multiclass learning using an ensemble-based cascaded learning framework. By implementing a multiclass cascaded classifier with AdaBoost, we show how detection runtimes are accelerated since only a subset of the ensemble is executed, thus making the classifiers suitable for computer vision applications. We also propose a new multiclass weak learner and demonstrate the framework's ability to achieve arbitrarily low training errors in conjunction with it. We tested our algorithm against AdaBoost.OC, ECC and M2 multiclass learning methods, on seven benchmark UCI datasets. In our experiments, we found that our framework achieves higher accuracy on five out of seven datasets and displays faster runtime efficiency in all cases.

UR - http://www.scopus.com/inward/record.url?scp=80052812680&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-23672-3_68

DO - 10.1007/978-3-642-23672-3_68

M3 - Conference contribution

AN - SCOPUS:80052812680

SN - 9783642236716

VL - PART 1

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

SP - 563

EP - 570

BT - Computer Analysis of Images and Patterns - 14th International Conference, CAIP 2011, Proceedings

A2 - Real, Pedro

A2 - Diaz-Pernil, Daniel

A2 - Molina-Abril, Helena

A2 - Berciano, Ainhoa

A2 - Kropatsch, Walter

PB - Springer

T2 - 14th International Conference on Computer Analysis of Images and Patterns, CAIP 2011

Y2 - 29 August 2011 through 31 August 2011

ER -

Susnjak T, Barczak A, Reyes N, Hawick K. A new ensemble-based cascaded framework for multiclass training with simple weak learners. In Real P, Diaz-Pernil D, Molina-Abril H, Berciano A, Kropatsch W, editors, Computer Analysis of Images and Patterns - 14th International Conference, CAIP 2011, Proceedings. Vol. PART 1. Springer. 2011. p. 563-570. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1). doi: 10.1007/978-3-642-23672-3_68

A new ensemble-based cascaded framework for multiclass training with simple weak learners

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this