Optimal linear feature selection for a general class of statistical pattern recognition models

A dimension reduction method proposed by Odell (1979) and Decell, Odell, and Coberly (1981) for Gaussian models is extended to a general class of density functions known as @q-generalized normal densities. Necessary and sufficient conditions for the existence of a dimension reduction matrix which preserves the original Bayes classification regions is derived. Moreover, an explicit expression for the compression matrix is given.

[1]  H. P. Decell,et al.  Linear dimension reduction and Bayes classification , 1981, Pattern Recognit..

[2]  S. Kotz,et al.  Multivariate θ-generalized normal distributions , 1973 .

[3]  Homer F. Walker,et al.  On Minimizing the Probability of Misclassification for Linear Feature Selection , 1975 .

[4]  Patrick L. Odell.  A model for dimension reduction in pattern recognition using continuous data , 1979, Pattern Recognit..

[5]  H. P. Decell,et al.  Linear feature selection with applications , 1979, Pattern Recognit..

[6]  Charles Peters.  Feature selection for best mean square approximation of class densities , 1979, Pattern Recognit..

[7]  Dean M. Young,et al.  A formulation and comparison of two linear feature selection techniques applicable to statistical classification , 1984, Pattern Recognit..

[8]  W. A. Coberly,et al.  Linear dimension reduction and Bayes classification with unknown population parameters , 1982, Pattern Recognit..

[9]  Henry P. Decell,et al.  Feature combinations and the bhattacharyya criterion , 1976 .

[10]  G. C. Tiao,et al.  The Use of OLUMV Estimators in Inference Robustness Studies of the Location Parameter of a Class of Symmetric Distributions , 1970 .

[11]  Raj S. Chhikara,et al.  Discriminant analysis using certain normed exponential densities with emphasis on remote sensing application , 1973, Pattern Recognit..

[12]  G. C. Tiao,et al.  A Further Look at Robustness via Bayes's Theorem , 1962 .

[13]  H. P. Decell,et al.  Feature combinations and the divergence criterion , 1977 .

[14]  L. F. Guseman,et al.  Distance preserving linear feature selection , 1979, Pattern Recognit..