Methods for Rapid Selection of Kernel Function Blur Coefficients in a Nonparametric Pattern Recognition Algorithm

A fast algorithm is proposed for choosing the coefficients of blur coefficients for kernel functions in a nonparametric estimate of the separating surface equation for a two-alternative pattern recognition problem. The algorithm is based on the results of a study of the asymptotic properties of nonp...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Measurement techniques Jg. 62; H. 4; S. 300 - 306
Hauptverfasser: Lapko, A. V., Lapko, V. A.
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York Springer US 01.07.2019
Springer
Springer Nature B.V
Schlagworte:
Algorithms
Analysis
Analytical Chemistry
Asymptotic properties
Characterization and Evaluation of Materials
Coefficients
Distribution (Probability theory)
General Problems of Metrology and Measurement Technique
Kernel functions
Measurement Science and Instrumentation
Methods
Nonparametric statistics
Pattern recognition
Physical Chemistry
Physics
Physics and Astronomy
Random variables
kernel probability density estimation
nonparametric pattern recognition algorithm
Rosenblatt-Parzen probability density estimate
quick selection of blur coefficients
ISSN:0543-1972, 1573-8906
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:A fast algorithm is proposed for choosing the coefficients of blur coefficients for kernel functions in a nonparametric estimate of the separating surface equation for a two-alternative pattern recognition problem. The algorithm is based on the results of a study of the asymptotic properties of nonparametric estimates of the decision function in the recognition problem for patterns and the probability densities of the distribution of random variables in classes. We compare the proposed algorithm with the traditional approach based on minimizing the estimated probability of a classification error.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0543-1972
1573-8906
DOI:10.1007/s11018-019-01621-1