Multiclass linear dimension reduction by weighted pairwise Fisher criteria

IEEE Transactions on Pattern Analysis and Machine Intelligence

Volumn 23, Issue 7, 2001, Pages 762-766

  Loog, Marco ^b   Duin, R P W ^a,c   Haeb Umbach, R ^a,d  

^a IEEE   (Germany)

^b UNIVERSITY MEDICAL CENTER UTRECHT   (Netherlands)

^c DELFT UNIVERSITY OF TECHNOLOGY   (Netherlands)

^d PHILIPS RESEARCH LABORATORIES   (Germany)

Author keywords

Approximate pairwise accuracy criterion; Bayes error; Fisher criterion; Linear dimension reduction; Linear discriminant analysis

Indexed keywords

LINEAR DIMENSION REDUCTION (LDR); LINEAR DISCRIMINANT ANALYSIS (LDA);

EIGENVALUES AND EIGENFUNCTIONS; MATHEMATICAL TRANSFORMATIONS; MATRIX ALGEBRA; MONTE CARLO METHODS; NEURAL NETWORKS; PROBABILISTIC LOGICS; TRANSFER FUNCTIONS;

PATTERN RECOGNITION;

EID: 0035393361     PISSN: 01628828     EISSN: None     Source Type: Journal    
DOI: 10.1109/34.935849     Document Type: Article

References (14)

1. Toward Bayes-optimal linear dimension reduction
2. Canonical variate analysis - A general model formulation
3. Feature combinations and the divergence criterion
4. Multi-class linear feature extraction by nonlinear PCA
5. The statistical utilization of multiple measurements
6. Discriminant analysis by Gaussian mixtures
7. Feature selection and extraction
8. Hetroscedastic discriminant analysis and reduced rank HMMs for improved speech recognition
9. The nonlinear PCA learning rule in independent component analysis
10. The utilization of multiple measurements in problems of biological classification
11. A formulation and comparison of two linear feature selection techniques applicable to statistical classification