Nearest neighbor method based on local distribution for classification

Chengsheng Mao; Bin Hu; Philip Moore; Yun Su; Manman Wang

doi:10.1007/978-3-319-18038-0_19

Nearest neighbor method based on local distribution for classification

Chengsheng Mao, Bin Hu^*, Philip Moore, Yun Su, Manman Wang

^*Corresponding author for this work

Lanzhou University

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

6 Citations (Scopus)

Abstract

The k-nearest-neighbor (kNN) algorithm is a simple but effective classification method which predicts the class label of a query sample based on information contained in its neighborhood. Previous versions of kNN usually consider the k nearest neighbors separately by the quantity or distance information. However, the quantity and the isolated distance information may be insufficient for effective classification decision. This paper investigates the kNN method from a perspective of local distribution based on which we propose an improved implementation of kNN. The proposed method performs the classification task by assigning the query sample to the class with the maximum posterior probability which is estimated from the local distribution based on the Bayesian rule. Experiments have been conducted using 15 benchmark datasets and the reported experimental results demonstrate excellent performance and robustness for the proposed method when compared to other state-of-the-art classifiers.

Original language	English
Title of host publication	Advances in Knowledge Discovery and Data Mining - 19th Pacific-Asia Conference, PAKDD 2015, Proceedings
Editors	Tu-Bao Ho, Hiroshi Motoda, Hiroshi Motoda, Ee-Peng Lim, Tru Cao, David Cheung, Zhi-Hua Zhou
Publisher	Springer Verlag
Pages	239-250
Number of pages	12
ISBN (Print)	9783319180373
DOIs	https://doi.org/10.1007/978-3-319-18038-0_19
Publication status	Published - 2015
Externally published	Yes
Event	19th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2015 - Ho Chi Minh City, Viet Nam Duration: 19 May 2015 → 22 May 2015

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9077
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	19th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2015
Country/Territory	Viet Nam
City	Ho Chi Minh City
Period	19/05/15 → 22/05/15

Keywords

Classification
Local distribution
Nearest neighbors
Posterior probability

Access to Document

10.1007/978-3-319-18038-0_19

Cite this

Mao, C., Hu, B., Moore, P., Su, Y., & Wang, M. (2015). Nearest neighbor method based on local distribution for classification. In T.-B. Ho, H. Motoda, H. Motoda, E.-P. Lim, T. Cao, D. Cheung, & Z.-H. Zhou (Eds.), Advances in Knowledge Discovery and Data Mining - 19th Pacific-Asia Conference, PAKDD 2015, Proceedings (pp. 239-250). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9077). Springer Verlag. https://doi.org/10.1007/978-3-319-18038-0_19

Mao, Chengsheng ; Hu, Bin ; Moore, Philip et al. / Nearest neighbor method based on local distribution for classification. Advances in Knowledge Discovery and Data Mining - 19th Pacific-Asia Conference, PAKDD 2015, Proceedings. editor / Tu-Bao Ho ; Hiroshi Motoda ; Hiroshi Motoda ; Ee-Peng Lim ; Tru Cao ; David Cheung ; Zhi-Hua Zhou. Springer Verlag, 2015. pp. 239-250 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{293d07e948834a7da84a5e911c729d2c,

title = "Nearest neighbor method based on local distribution for classification",

abstract = "The k-nearest-neighbor (kNN) algorithm is a simple but effective classification method which predicts the class label of a query sample based on information contained in its neighborhood. Previous versions of kNN usually consider the k nearest neighbors separately by the quantity or distance information. However, the quantity and the isolated distance information may be insufficient for effective classification decision. This paper investigates the kNN method from a perspective of local distribution based on which we propose an improved implementation of kNN. The proposed method performs the classification task by assigning the query sample to the class with the maximum posterior probability which is estimated from the local distribution based on the Bayesian rule. Experiments have been conducted using 15 benchmark datasets and the reported experimental results demonstrate excellent performance and robustness for the proposed method when compared to other state-of-the-art classifiers.",

keywords = "Classification, Local distribution, Nearest neighbors, Posterior probability",

author = "Chengsheng Mao and Bin Hu and Philip Moore and Yun Su and Manman Wang",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2015.; 19th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2015 ; Conference date: 19-05-2015 Through 22-05-2015",

year = "2015",

doi = "10.1007/978-3-319-18038-0_19",

language = "English",

isbn = "9783319180373",

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

publisher = "Springer Verlag",

pages = "239--250",

editor = "Tu-Bao Ho and Hiroshi Motoda and Hiroshi Motoda and Ee-Peng Lim and Tru Cao and David Cheung and Zhi-Hua Zhou",

booktitle = "Advances in Knowledge Discovery and Data Mining - 19th Pacific-Asia Conference, PAKDD 2015, Proceedings",

address = "Germany",

}

Mao, C, Hu, B, Moore, P, Su, Y & Wang, M 2015, Nearest neighbor method based on local distribution for classification. in T-B Ho, H Motoda, H Motoda, E-P Lim, T Cao, D Cheung & Z-H Zhou (eds), Advances in Knowledge Discovery and Data Mining - 19th Pacific-Asia Conference, PAKDD 2015, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9077, Springer Verlag, pp. 239-250, 19th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2015, Ho Chi Minh City, Viet Nam, 19/05/15. https://doi.org/10.1007/978-3-319-18038-0_19

Nearest neighbor method based on local distribution for classification. / Mao, Chengsheng; Hu, Bin; Moore, Philip et al.
Advances in Knowledge Discovery and Data Mining - 19th Pacific-Asia Conference, PAKDD 2015, Proceedings. ed. / Tu-Bao Ho; Hiroshi Motoda; Hiroshi Motoda; Ee-Peng Lim; Tru Cao; David Cheung; Zhi-Hua Zhou. Springer Verlag, 2015. p. 239-250 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9077).

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

TY - GEN

T1 - Nearest neighbor method based on local distribution for classification

AU - Mao, Chengsheng

AU - Hu, Bin

AU - Moore, Philip

AU - Su, Yun

AU - Wang, Manman

N1 - Publisher Copyright: © Springer International Publishing Switzerland 2015.

PY - 2015

Y1 - 2015

N2 - The k-nearest-neighbor (kNN) algorithm is a simple but effective classification method which predicts the class label of a query sample based on information contained in its neighborhood. Previous versions of kNN usually consider the k nearest neighbors separately by the quantity or distance information. However, the quantity and the isolated distance information may be insufficient for effective classification decision. This paper investigates the kNN method from a perspective of local distribution based on which we propose an improved implementation of kNN. The proposed method performs the classification task by assigning the query sample to the class with the maximum posterior probability which is estimated from the local distribution based on the Bayesian rule. Experiments have been conducted using 15 benchmark datasets and the reported experimental results demonstrate excellent performance and robustness for the proposed method when compared to other state-of-the-art classifiers.

AB - The k-nearest-neighbor (kNN) algorithm is a simple but effective classification method which predicts the class label of a query sample based on information contained in its neighborhood. Previous versions of kNN usually consider the k nearest neighbors separately by the quantity or distance information. However, the quantity and the isolated distance information may be insufficient for effective classification decision. This paper investigates the kNN method from a perspective of local distribution based on which we propose an improved implementation of kNN. The proposed method performs the classification task by assigning the query sample to the class with the maximum posterior probability which is estimated from the local distribution based on the Bayesian rule. Experiments have been conducted using 15 benchmark datasets and the reported experimental results demonstrate excellent performance and robustness for the proposed method when compared to other state-of-the-art classifiers.

KW - Classification

KW - Local distribution

KW - Nearest neighbors

KW - Posterior probability

UR - http://www.scopus.com/inward/record.url?scp=84945966441&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-18038-0_19

DO - 10.1007/978-3-319-18038-0_19

M3 - Conference contribution

AN - SCOPUS:84945966441

SN - 9783319180373

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

SP - 239

EP - 250

BT - Advances in Knowledge Discovery and Data Mining - 19th Pacific-Asia Conference, PAKDD 2015, Proceedings

A2 - Ho, Tu-Bao

A2 - Motoda, Hiroshi

A2 - Lim, Ee-Peng

A2 - Cao, Tru

A2 - Cheung, David

A2 - Zhou, Zhi-Hua

PB - Springer Verlag

T2 - 19th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2015

Y2 - 19 May 2015 through 22 May 2015

ER -

Mao C, Hu B, Moore P, Su Y, Wang M. Nearest neighbor method based on local distribution for classification. In Ho TB, Motoda H, Motoda H, Lim EP, Cao T, Cheung D, Zhou ZH, editors, Advances in Knowledge Discovery and Data Mining - 19th Pacific-Asia Conference, PAKDD 2015, Proceedings. Springer Verlag. 2015. p. 239-250. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-18038-0_19

Nearest neighbor method based on local distribution for classification

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this