Methods for Rapid Selection of Kernel Function Blur Coefficients in a Nonparametric Pattern Recognition Algorithm
- Authors: Lapko A.V.1,2, Lapko V.A.1,2
-
Affiliations:
- Institute of Computational Modeling, Siberian Branch of the Russian Academy of Sciences
- Siberian State University of Science and Technology
- Issue: Vol 62, No 4 (2019)
- Pages: 300-306
- Section: General Problems of Metrology and Measurement Technique
- URL: https://journals.rcsi.science/0543-1972/article/view/246705
- DOI: https://doi.org/10.1007/s11018-019-01621-1
- ID: 246705
Cite item
Abstract
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.
About the authors
A. V. Lapko
Institute of Computational Modeling, Siberian Branch of the Russian Academy of Sciences; Siberian State University of Science and Technology
Author for correspondence.
Email: lapko@icm.krasn.ru
Russian Federation, Krasnoyarsk; Krasnoyarsk
V. A. Lapko
Institute of Computational Modeling, Siberian Branch of the Russian Academy of Sciences; Siberian State University of Science and Technology
Email: lapko@icm.krasn.ru
Russian Federation, Krasnoyarsk; Krasnoyarsk