TY - JOUR
T1 - Second-order measures of quality for binary classification
T2 - A critical overview and their use for nonlinear receiver design
AU - Abdallah, Fahed
AU - Richard, Cédric
AU - Lengellé, Régis
PY - 2003/10
Y1 - 2003/10
N2 - When deriving a detector, we are often led to consider design criteria such as second-order measures of quality. The aim of this paper is to provide a critical overview of these criteria. We first consider the case of deriving unconstrained detectors. We show that second-order criteria must satisfy a non-trivial condition to yield Bayes-optimal receivers, to be considered as relevant criteria for detector design. Next, we address the case where constraints are imposed on the detection structure, leading us to consider some set D of admissible detectors. In these conditions we prove that even if there exists a monotonic function of the likelihood ratio in D obtaining this statistic via the optimization of a second-order criterion, relevant or not, is not guaranteed. Results are illustrated by simulation examples. Finally, in order to derive nonlinear discriminants via optimization of second-order criteria, we propose a method based on the kernel trick used in the implementation of the well-known support vector machine method. The new method is tested on a number of real data sets.
AB - When deriving a detector, we are often led to consider design criteria such as second-order measures of quality. The aim of this paper is to provide a critical overview of these criteria. We first consider the case of deriving unconstrained detectors. We show that second-order criteria must satisfy a non-trivial condition to yield Bayes-optimal receivers, to be considered as relevant criteria for detector design. Next, we address the case where constraints are imposed on the detection structure, leading us to consider some set D of admissible detectors. In these conditions we prove that even if there exists a monotonic function of the likelihood ratio in D obtaining this statistic via the optimization of a second-order criterion, relevant or not, is not guaranteed. Results are illustrated by simulation examples. Finally, in order to derive nonlinear discriminants via optimization of second-order criteria, we propose a method based on the kernel trick used in the implementation of the well-known support vector machine method. The new method is tested on a number of real data sets.
KW - Detection
KW - Distance measures
KW - Maximum likelihood
KW - Nonlinear discriminants
KW - Signal-to-noise ratio
UR - http://www.scopus.com/inward/record.url?scp=3042670454&partnerID=8YFLogxK
U2 - 10.1080/10255810390243700
DO - 10.1080/10255810390243700
M3 - Article
AN - SCOPUS:3042670454
SN - 1025-5818
VL - 5
SP - 401
EP - 408
JO - International Journal of Smart Engineering System Design
JF - International Journal of Smart Engineering System Design
IS - 4
ER -