Rectified nearest feature line segment for pattern classification

被引：30

作者：

Du, Hao ^{[1
]}

Chen, Yan Qiu ^{[1
]}

机构：

[1] Fudan Univ, Dept Comp Sci & Engn, Sch Informat Sci & Engn, Shanghai 200433, Peoples R China

来源：

PATTERN RECOGNITION | 2007年 / 40卷 / 05期

基金：

中国国家自然科学基金;

关键词：

pattern classification; nearest feature line; rectified nearest feature line segment; distribution concentration; interpolation and extrapolation accuracy;

D O I：

10.1016/j.patcog.2006.10.021

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This paper points out and analyzes the advantages and drawbacks of the nearest feature line (NFL) classifier. To overcome the shortcomings, a new feature subspace with two simple and effective improvements is built to represent each class. The proposed method, termed rectified nearest feature line segment (RNFLS), is shown to possess a novel property of concentration as a result of the added line segments (features), which significantly enhances the classification ability. Another remarkable merit is that RNFLS is applicable to complex tasks such as the two-spiral distribution, which the original NFL cannot deal with properly. Finally, experimental comparisons with NFL, NN(nearest neighbor), k-NN and NNL (nearest neighbor line) using both artificial and real-world data-sets demonstrate that RNFLS offers the best performance. (c) 2006 Pattern Recognition Society. Published by Elsevier Ltd. All rights reserved.

引用

页码：1486 / 1497

页数：12

共 15 条

[1]

Blake C.L., 1998, UCI repository of machine learning databases

[2] Object recognition based on image sequences by using inter-feature-line consistencies [J].

Chen, JH ;

Chen, CS .

PATTERN RECOGNITION, 2004, 37 (09) :1913-1923

[3] On the use of nearest feature line for speaker identification [J].

Chen, K ;

Wu, TY ;

Zhang, HJ .

PATTERN RECOGNITION LETTERS, 2002, 23 (14) :1735-1746

[4]

CHEN YQ, 1994, NEURAL NETWORKS, V7, P1477, DOI 10.1016/0893-6080(94)90093-0

[5] Discriminant waveletfaces and nearest feature classifiers for face recognition [J].

Chien, JT ;

Wu, CC .

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2002, 24 (12) :1644-1649

[6] INITIALIZING BACK PROPAGATION NETWORKS WITH PROTOTYPES [J].

DENOEUX, T ;

LENGELLE, R .

NEURAL NETWORKS, 1993, 6 (03) :351-363

[7] LEAVE-ONE-OUT PROCEDURES FOR NONPARAMETRIC ERROR-ESTIMATES [J].

FUKUNAGA, K ;

HUMMELS, DM .

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1989, 11 (04) :421-423

[8] Learning pattern classification - A survey [J].

Kulkarni, SR ;

Lugosi, G ;

Venkatesh, SS .

IEEE TRANSACTIONS ON INFORMATION THEORY, 1998, 44 (06) :2178-2206

[9]

Li SZ, 2000, IEEE T PATTERN ANAL, V22, P1335, DOI 10.1109/34.888719

[10] Face recognition using the nearest feature line method [J].

Li, SZ ;

Lu, JW .

IEEE TRANSACTIONS ON NEURAL NETWORKS, 1999, 10 (02) :439-443

← 1 2 →