Fusion of pairwise nearest-neighbor classifiers based on pairwise-weighted distance metric and Dempster-Shafer theory

Lianmeng Jiao; Thierry Denoux; Quan Pan

Fusion of pairwise nearest-neighbor classifiers based on pairwise-weighted distance metric and Dempster-Shafer theory

Lianmeng Jiao, Thierry Denoux, Quan Pan

School of Automation

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

3 Scopus citations

Abstract

The performance of the nearest-neighbor (NN) classifier is known to be very sensitive to the distance metric used in classifying a query pattern, especially in scarce-prototype cases. In this paper, a pairwise-weighted (PW) distance metric related to pairs of class labels is proposed. Compared with the existing distance metrics, it provides more flexibility to design the feature weights so that the local specifics in feature space can be well characterized. Base on the proposed PW distance metric, a polychotomous NN classification problem is solved by combining several pairwise NN (PNN) classifiers within the framework of Dempster-Shafer theory to deal with the uncertain output information. Two experiments based on synthetic and real data sets were carried out to show the effectiveness of the proposed method.

Original language	English
Title of host publication	FUSION 2014 - 17th International Conference on Information Fusion
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9788490123553
State	Published - 3 Oct 2014
Event	17th International Conference on Information Fusion, FUSION 2014 - Salamanca, Spain Duration: 7 Jul 2014 → 10 Jul 2014

Publication series

Name	FUSION 2014 - 17th International Conference on Information Fusion

Conference

Conference	17th International Conference on Information Fusion, FUSION 2014
Country/Territory	Spain
City	Salamanca
Period	7/07/14 → 10/07/14

Keywords

Dempster-Shafer theory
nearest-neighbor classifier
pairwise-weighted distance metric
pattern classification

Cite this

@inproceedings{71b0d5e4e88f498e8a1dc2262f8c42fe,

title = "Fusion of pairwise nearest-neighbor classifiers based on pairwise-weighted distance metric and Dempster-Shafer theory",

abstract = "The performance of the nearest-neighbor (NN) classifier is known to be very sensitive to the distance metric used in classifying a query pattern, especially in scarce-prototype cases. In this paper, a pairwise-weighted (PW) distance metric related to pairs of class labels is proposed. Compared with the existing distance metrics, it provides more flexibility to design the feature weights so that the local specifics in feature space can be well characterized. Base on the proposed PW distance metric, a polychotomous NN classification problem is solved by combining several pairwise NN (PNN) classifiers within the framework of Dempster-Shafer theory to deal with the uncertain output information. Two experiments based on synthetic and real data sets were carried out to show the effectiveness of the proposed method.",

keywords = "Dempster-Shafer theory, nearest-neighbor classifier, pairwise-weighted distance metric, pattern classification",

author = "Lianmeng Jiao and Thierry Denoux and Quan Pan",

note = "Publisher Copyright: {\textcopyright} 2014 International Society of Information Fusion.; 17th International Conference on Information Fusion, FUSION 2014 ; Conference date: 07-07-2014 Through 10-07-2014",

year = "2014",

month = oct,

day = "3",

language = "英语",

series = "FUSION 2014 - 17th International Conference on Information Fusion",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "FUSION 2014 - 17th International Conference on Information Fusion",

}

Jiao, L, Denoux, T & Pan, Q 2014, Fusion of pairwise nearest-neighbor classifiers based on pairwise-weighted distance metric and Dempster-Shafer theory. in FUSION 2014 - 17th International Conference on Information Fusion., 6916087, FUSION 2014 - 17th International Conference on Information Fusion, Institute of Electrical and Electronics Engineers Inc., 17th International Conference on Information Fusion, FUSION 2014, Salamanca, Spain, 7/07/14.

Fusion of pairwise nearest-neighbor classifiers based on pairwise-weighted distance metric and Dempster-Shafer theory. / Jiao, Lianmeng; Denoux, Thierry; Pan, Quan.
FUSION 2014 - 17th International Conference on Information Fusion. Institute of Electrical and Electronics Engineers Inc., 2014. 6916087 (FUSION 2014 - 17th International Conference on Information Fusion).

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

TY - GEN

T1 - Fusion of pairwise nearest-neighbor classifiers based on pairwise-weighted distance metric and Dempster-Shafer theory

AU - Jiao, Lianmeng

AU - Denoux, Thierry

AU - Pan, Quan

PY - 2014/10/3

Y1 - 2014/10/3

N2 - The performance of the nearest-neighbor (NN) classifier is known to be very sensitive to the distance metric used in classifying a query pattern, especially in scarce-prototype cases. In this paper, a pairwise-weighted (PW) distance metric related to pairs of class labels is proposed. Compared with the existing distance metrics, it provides more flexibility to design the feature weights so that the local specifics in feature space can be well characterized. Base on the proposed PW distance metric, a polychotomous NN classification problem is solved by combining several pairwise NN (PNN) classifiers within the framework of Dempster-Shafer theory to deal with the uncertain output information. Two experiments based on synthetic and real data sets were carried out to show the effectiveness of the proposed method.

AB - The performance of the nearest-neighbor (NN) classifier is known to be very sensitive to the distance metric used in classifying a query pattern, especially in scarce-prototype cases. In this paper, a pairwise-weighted (PW) distance metric related to pairs of class labels is proposed. Compared with the existing distance metrics, it provides more flexibility to design the feature weights so that the local specifics in feature space can be well characterized. Base on the proposed PW distance metric, a polychotomous NN classification problem is solved by combining several pairwise NN (PNN) classifiers within the framework of Dempster-Shafer theory to deal with the uncertain output information. Two experiments based on synthetic and real data sets were carried out to show the effectiveness of the proposed method.

KW - Dempster-Shafer theory

KW - nearest-neighbor classifier

KW - pairwise-weighted distance metric

KW - pattern classification

UR - http://www.scopus.com/inward/record.url?scp=84910599765&partnerID=8YFLogxK

M3 - 会议稿件

AN - SCOPUS:84910599765

T3 - FUSION 2014 - 17th International Conference on Information Fusion

BT - FUSION 2014 - 17th International Conference on Information Fusion

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 17th International Conference on Information Fusion, FUSION 2014

Y2 - 7 July 2014 through 10 July 2014

ER -

Fusion of pairwise nearest-neighbor classifiers based on pairwise-weighted distance metric and Dempster-Shafer theory

Abstract

Publication series

Conference

Keywords

Other files and links

Fingerprint

Cite this