Article

Article title DECISION RULE’S SYNTHESIS OF SIGNAL CLASSIFIER IN DISTRIBUTION-FREE PRIOR UNCERTAINTY ENVIRONMENT
Authors G.G. Galustov, A.A. Potsykaylo, D.A. Krasnobayev
Section SECTION II. PROCESSING OF SIGNALS
Month, Year 01, 2011 @en
Index UDC
DOI
Abstract There is review method of decision rule’s synthesis embodied based on approximating approach formation of plausibility function according clustered sample in distribution-free prior uncertainty environment. It is shown for example is algorithm of determination of an estimation of a coefficient vector for a Fourier-series expansion of plausibility function on orthonormal system. It allows to discover an estimation of the plausibility function which use in a solving rule. The block diagramme of the distinguishing device of signals is shown at the simple loss function, constructed on criterion of a minimum for average probability of a recognition error. Thus the minimum average error of recognition is ensured with use of the ideal observer criterion in case of classification of the signals characterised by one indication with univariate non-Gauissian distribution and at unknown probabilities of recognition of classes.

Download PDF

Keywords Decision rule; approximating approach; function according; clustered sample; distribution-free prior uncertainty environment.
References 1. Ту Дж., Гонсалес Р. Принципы распознавания образов. – М.: Мир, 1978. – 411 с.
2. Омельченко В.А. Распознавание сигналов по спектру в условиях априорной неопределенности. – Харьков: ХПИ, 1979. – 100 с.
3. Фукунага К. Введение в статистическую теорию распознавания образов. – М.: Наука, 1979. – 367 с.
4. Галустов Г.Г. Классификатор случайных сигналов // Известия СКНЦ ВШ, серия
Технические науки. – 1984. – № 3. – С. 6.
5. Галустов Г.Г. Моделирование случайных процессов и оценивание их статистиче-
ских характеристик. – М.: Радио и связь, 1999. – 120 с.

Comments are closed.