A combined classification algorithm based on C4.5 and NB

Liangxiao Jiang; Chaoqun Li; Jia Wu; Jian Zhu

doi:10.1007/978-3-540-92137-0_39

A combined classification algorithm based on C4.5 and NB

Liangxiao Jiang, Chaoqun Li, Jia Wu, Jian Zhu

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

7 Citations (Scopus)

Abstract

When our learning task is to build a model with accurate classification, C4.5 and NB are two very important algorithms for achieving this task because of their simplicity and high performance. In this paper, we present a combined classification algorithm based on C4.5 and NB, simply C4.5-NB. In C4.5-NB, the class probability estimates of C4.5 and NB are weighted according to their classification accuracy on the training data. We experimentally tested C4.5-NB in Weka system using the whole 36 UCI data sets selected by Weka, and compared it with C4.5 and NB. The experimental results show that C4.5-NB significantly outperforms C4.5 and NB in terms of classification accuracy. Besides, we also observe the ranking performance of C4.5-NB in terms of AUC (the area under the Receiver Operating Characteristics curve). Fortunately, C4.5-NB also significantly outperforms C4.5 and NB.

Original language	English
Title of host publication	Advances in Computation and Intelligence
Subtitle of host publication	Third International Symposium, ISICA 2008, Proceedings
Editors	Lishan Kang, Zhihua Cai, Xuesong Yan, Yong Liu
Place of Publication	Berlin
Publisher	Springer, Springer Nature
Pages	350-359
Number of pages	10
ISBN (Electronic)	9783540921370
ISBN (Print)	9783540921363
DOIs	https://doi.org/10.1007/978-3-540-92137-0_39
Publication status	Published - 2008
Externally published	Yes
Event	3rd International Symposium on Intelligence Computation and Applications (ISICA 2008) - Wuhan, China Duration: 19 Dec 2008 → 21 Dec 2008

Conference

Conference	3rd International Symposium on Intelligence Computation and Applications (ISICA 2008)
Country/Territory	China
City	Wuhan
Period	19/12/08 → 21/12/08

Keywords

decision trees
naive Bayes
combined algorithms
weights
classification
ranking
data mining

Access to Document

10.1007/978-3-540-92137-0_39

Cite this

@inproceedings{5c01b36d53cb4cb4bdaac22cf6a71d41,

title = "A combined classification algorithm based on C4.5 and NB",

abstract = "When our learning task is to build a model with accurate classification, C4.5 and NB are two very important algorithms for achieving this task because of their simplicity and high performance. In this paper, we present a combined classification algorithm based on C4.5 and NB, simply C4.5-NB. In C4.5-NB, the class probability estimates of C4.5 and NB are weighted according to their classification accuracy on the training data. We experimentally tested C4.5-NB in Weka system using the whole 36 UCI data sets selected by Weka, and compared it with C4.5 and NB. The experimental results show that C4.5-NB significantly outperforms C4.5 and NB in terms of classification accuracy. Besides, we also observe the ranking performance of C4.5-NB in terms of AUC (the area under the Receiver Operating Characteristics curve). Fortunately, C4.5-NB also significantly outperforms C4.5 and NB.",

keywords = "decision trees, naive Bayes, combined algorithms, weights, classification, ranking, data mining",

author = "Liangxiao Jiang and Chaoqun Li and Jia Wu and Jian Zhu",

year = "2008",

doi = "10.1007/978-3-540-92137-0_39",

language = "English",

isbn = "9783540921363",

pages = "350--359",

editor = "Lishan Kang and Zhihua Cai and Xuesong Yan and Yong Liu",

booktitle = "Advances in Computation and Intelligence",

publisher = "Springer, Springer Nature",

address = "United States",

note = "3rd International Symposium on Intelligence Computation and Applications (ISICA 2008) ; Conference date: 19-12-2008 Through 21-12-2008",

}

Jiang, L, Li, C, Wu, J & Zhu, J 2008, A combined classification algorithm based on C4.5 and NB. in L Kang, Z Cai, X Yan & Y Liu (eds), Advances in Computation and Intelligence: Third International Symposium, ISICA 2008, Proceedings. Springer, Springer Nature, Berlin, pp. 350-359, 3rd International Symposium on Intelligence Computation and Applications (ISICA 2008), Wuhan, China, 19/12/08. https://doi.org/10.1007/978-3-540-92137-0_39

A combined classification algorithm based on C4.5 and NB. / Jiang, Liangxiao; Li, Chaoqun; Wu, Jia et al.
Advances in Computation and Intelligence: Third International Symposium, ISICA 2008, Proceedings. ed. / Lishan Kang; Zhihua Cai; Xuesong Yan; Yong Liu. Berlin: Springer, Springer Nature, 2008. p. 350-359.

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

TY - GEN

T1 - A combined classification algorithm based on C4.5 and NB

AU - Jiang, Liangxiao

AU - Li, Chaoqun

AU - Wu, Jia

AU - Zhu, Jian

PY - 2008

Y1 - 2008

N2 - When our learning task is to build a model with accurate classification, C4.5 and NB are two very important algorithms for achieving this task because of their simplicity and high performance. In this paper, we present a combined classification algorithm based on C4.5 and NB, simply C4.5-NB. In C4.5-NB, the class probability estimates of C4.5 and NB are weighted according to their classification accuracy on the training data. We experimentally tested C4.5-NB in Weka system using the whole 36 UCI data sets selected by Weka, and compared it with C4.5 and NB. The experimental results show that C4.5-NB significantly outperforms C4.5 and NB in terms of classification accuracy. Besides, we also observe the ranking performance of C4.5-NB in terms of AUC (the area under the Receiver Operating Characteristics curve). Fortunately, C4.5-NB also significantly outperforms C4.5 and NB.

AB - When our learning task is to build a model with accurate classification, C4.5 and NB are two very important algorithms for achieving this task because of their simplicity and high performance. In this paper, we present a combined classification algorithm based on C4.5 and NB, simply C4.5-NB. In C4.5-NB, the class probability estimates of C4.5 and NB are weighted according to their classification accuracy on the training data. We experimentally tested C4.5-NB in Weka system using the whole 36 UCI data sets selected by Weka, and compared it with C4.5 and NB. The experimental results show that C4.5-NB significantly outperforms C4.5 and NB in terms of classification accuracy. Besides, we also observe the ranking performance of C4.5-NB in terms of AUC (the area under the Receiver Operating Characteristics curve). Fortunately, C4.5-NB also significantly outperforms C4.5 and NB.

KW - decision trees

KW - naive Bayes

KW - combined algorithms

KW - weights

KW - classification

KW - ranking

KW - data mining

U2 - 10.1007/978-3-540-92137-0_39

DO - 10.1007/978-3-540-92137-0_39

M3 - Conference proceeding contribution

SN - 9783540921363

SP - 350

EP - 359

BT - Advances in Computation and Intelligence

A2 - Kang, Lishan

A2 - Cai, Zhihua

A2 - Yan, Xuesong

A2 - Liu, Yong

PB - Springer, Springer Nature

CY - Berlin

T2 - 3rd International Symposium on Intelligence Computation and Applications (ISICA 2008)

Y2 - 19 December 2008 through 21 December 2008

ER -