Contributions of domain knowledge and stacked generalization in AI-based classification models

Weiping Kostenko; Vincent Lee; Ting Tan

doi:10.1007/978-3-540-30549-1_100

Contributions of domain knowledge and stacked generalization in AI-based classification models

Weiping Kostenko, Vincent Lee, Ting Tan

Econometrics & Business Statistics

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

Abstract

We exploit the merits of C4.5 decision tree classifier with two stacking meta-learners: back-propagation multilayer perceptron neural network and naive-Bayes respectively. The performance of these two hybrid classification schemes have been empirically tested and compared with C4.5 decision tree using two US data sets (raw data set and new data set incorporated with domain knowledge) simultaneously to predict US bank failure. Significant improvements in prediction accuracy and training efficiency have been achieved in the schemes based on new data set. The empirical test results suggest that the proposed hybrid schemes perform marginally better in term of AUC criterion.

Original language	English
Title of host publication	AI 2004: Advances in Artificial Intelligence
Subtitle of host publication	17th Australian Joint Conference on Artificial Intelligence Cairns, Australia, December 4-6, 2004 Proceedings
Editors	Geoffrey I. Webb, Xinghuo Yu
Place of Publication	Berlin Germany
Publisher	Springer
Pages	1049-1054
Number of pages	6
ISBN (Print)	3540240594
DOIs	https://doi.org/10.1007/978-3-540-30549-1_100
Publication status	Published - 2004
Event	Australasian Joint Conference on Artificial Intelligence 2004 - Cairns, Australia Duration: 4 Dec 2004 → 6 Dec 2004 Conference number: 17th https://link.springer.com/book/10.1007/b104336 (Proceedings)

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer
Volume	3339
ISSN (Print)	0302-9743

Conference

Conference	Australasian Joint Conference on Artificial Intelligence 2004
Abbreviated title	AI 2004
Country/Territory	Australia
City	Cairns
Period	4/12/04 → 6/12/04
Internet address	https://link.springer.com/book/10.1007/b104336 (Proceedings)

Access to Document

10.1007/978-3-540-30549-1_100

Cite this

Kostenko, W., Lee, V., & Tan, T. (2004). Contributions of domain knowledge and stacked generalization in AI-based classification models. In G. I. Webb, & X. Yu (Eds.), AI 2004: Advances in Artificial Intelligence: 17th Australian Joint Conference on Artificial Intelligence Cairns, Australia, December 4-6, 2004 Proceedings (pp. 1049-1054). (Lecture Notes in Computer Science; Vol. 3339). Springer. https://doi.org/10.1007/978-3-540-30549-1_100

Kostenko, Weiping ; Lee, Vincent ; Tan, Ting. / Contributions of domain knowledge and stacked generalization in AI-based classification models. AI 2004: Advances in Artificial Intelligence: 17th Australian Joint Conference on Artificial Intelligence Cairns, Australia, December 4-6, 2004 Proceedings. editor / Geoffrey I. Webb ; Xinghuo Yu. Berlin Germany : Springer, 2004. pp. 1049-1054 (Lecture Notes in Computer Science).

@inproceedings{3844e34e117746a2aed9cc1a4b0a1b2d,

title = "Contributions of domain knowledge and stacked generalization in AI-based classification models",

abstract = "We exploit the merits of C4.5 decision tree classifier with two stacking meta-learners: back-propagation multilayer perceptron neural network and naive-Bayes respectively. The performance of these two hybrid classification schemes have been empirically tested and compared with C4.5 decision tree using two US data sets (raw data set and new data set incorporated with domain knowledge) simultaneously to predict US bank failure. Significant improvements in prediction accuracy and training efficiency have been achieved in the schemes based on new data set. The empirical test results suggest that the proposed hybrid schemes perform marginally better in term of AUC criterion.",

author = "Weiping Kostenko and Vincent Lee and Ting Tan",

year = "2004",

doi = "10.1007/978-3-540-30549-1_100",

language = "English",

isbn = "3540240594",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "1049--1054",

editor = "Webb, {Geoffrey I.} and Xinghuo Yu",

booktitle = "AI 2004: Advances in Artificial Intelligence",

note = "Australasian Joint Conference on Artificial Intelligence 2004, AI 2004 ; Conference date: 04-12-2004 Through 06-12-2004",

url = "https://link.springer.com/book/10.1007/b104336",

}

Kostenko, W, Lee, V & Tan, T 2004, Contributions of domain knowledge and stacked generalization in AI-based classification models. in GI Webb & X Yu (eds), AI 2004: Advances in Artificial Intelligence: 17th Australian Joint Conference on Artificial Intelligence Cairns, Australia, December 4-6, 2004 Proceedings. Lecture Notes in Computer Science, vol. 3339, Springer, Berlin Germany, pp. 1049-1054, Australasian Joint Conference on Artificial Intelligence 2004, Cairns, Queensland, Australia, 4/12/04. https://doi.org/10.1007/978-3-540-30549-1_100

Contributions of domain knowledge and stacked generalization in AI-based classification models. / Kostenko, Weiping; Lee, Vincent; Tan, Ting.
AI 2004: Advances in Artificial Intelligence: 17th Australian Joint Conference on Artificial Intelligence Cairns, Australia, December 4-6, 2004 Proceedings. ed. / Geoffrey I. Webb; Xinghuo Yu. Berlin Germany: Springer, 2004. p. 1049-1054 (Lecture Notes in Computer Science; Vol. 3339).

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

TY - GEN

T1 - Contributions of domain knowledge and stacked generalization in AI-based classification models

AU - Kostenko, Weiping

AU - Lee, Vincent

AU - Tan, Ting

N1 - Conference code: 17th

PY - 2004

Y1 - 2004

N2 - We exploit the merits of C4.5 decision tree classifier with two stacking meta-learners: back-propagation multilayer perceptron neural network and naive-Bayes respectively. The performance of these two hybrid classification schemes have been empirically tested and compared with C4.5 decision tree using two US data sets (raw data set and new data set incorporated with domain knowledge) simultaneously to predict US bank failure. Significant improvements in prediction accuracy and training efficiency have been achieved in the schemes based on new data set. The empirical test results suggest that the proposed hybrid schemes perform marginally better in term of AUC criterion.

AB - We exploit the merits of C4.5 decision tree classifier with two stacking meta-learners: back-propagation multilayer perceptron neural network and naive-Bayes respectively. The performance of these two hybrid classification schemes have been empirically tested and compared with C4.5 decision tree using two US data sets (raw data set and new data set incorporated with domain knowledge) simultaneously to predict US bank failure. Significant improvements in prediction accuracy and training efficiency have been achieved in the schemes based on new data set. The empirical test results suggest that the proposed hybrid schemes perform marginally better in term of AUC criterion.

U2 - 10.1007/978-3-540-30549-1_100

DO - 10.1007/978-3-540-30549-1_100

M3 - Conference Paper

SN - 3540240594

T3 - Lecture Notes in Computer Science

SP - 1049

EP - 1054

BT - AI 2004: Advances in Artificial Intelligence

A2 - Webb, Geoffrey I.

A2 - Yu, Xinghuo

PB - Springer

CY - Berlin Germany

T2 - Australasian Joint Conference on Artificial Intelligence 2004

Y2 - 4 December 2004 through 6 December 2004

ER -

Kostenko W, Lee V, Tan T. Contributions of domain knowledge and stacked generalization in AI-based classification models. In Webb GI, Yu X, editors, AI 2004: Advances in Artificial Intelligence: 17th Australian Joint Conference on Artificial Intelligence Cairns, Australia, December 4-6, 2004 Proceedings. Berlin Germany: Springer. 2004. p. 1049-1054. (Lecture Notes in Computer Science). doi: 10.1007/978-3-540-30549-1_100