Histogram Filter with Smoothing Parameter Setting
https://doi.org/10.35596/1729-7648-2022-20-8-42-50
Abstract
A histogram filter with smoothing parameter settings is discussed in the article. The histogram filter can be effectively applied in the problems of identification (recognition) of distribution laws for small amounts of data. The smoothing parameter is determined taking into account the available a priori information regarding the proposed distribution law. The relationship between the mathematical expectations of the chi-square fit criterion of the standard estimation histogram and the use of the histogram filter has been determined. This ratio is determined by the smoothing factor. The numerical value of the smoothing coefficient depends on the following parameters: the amount of data, the number of grouping intervals, and the shape parameters of the distribution law. The paper analyzes the feasibility of using a histogram filter, depending on the ratio of the above parameters. The dependence of the smoothing coefficient on the specified parameters allows one to determine the relationship between the number of data grouping intervals and their volume. The histogram filter is an easy-to-implement tool that can be easily integrated into any open distribution law identification (recognition) algorithm
About the Authors
A. V. Ausiannikau
Belarusian State University
Belarus
Belarus
Ausiannikau Andrei Vital’evich, Cand. of Sci., Assistant Professor,
Assistant Professor at the Department of Information Technologies
220030, Minsk, Nezavisimosti Ave., 4
Tel. +375 17 209-58-94
V. M. Kozel
Belarusian State University of Informatics and Radioelectronics
Belarus
Belarus
Kozel V. M., Cand. of Sci., Assistant Professor, Assistant Professor at the Department of Information Radiotechnologies
Minsk
References
1. Orlov Yu. N. (2013) Optimal Histogram Partitioning for Estimating the Sample Density of the Distribution Function of an Unsteady Time Series. Preprinty IPM imeni M. V. Keldysha =Preprints of the Institute of Applied Mathematics name after M. V. Keldysh RAS. (14), 26. http://library.keldysh.ru/preprint.asp?id=2013-14.
2. Chong Gu, Yongho Jeon, Yi Lin (2013) Nonparametric Density Estimation in High-Dimensions. Statistica Sinica, 23, 1131–1153.
3. Devroye L., Gyorfi L. (1985) Nonparametric Density Estimation: the L1 View. New York, John Wiley Sons Publ.
4. Solomon C. J., Breckon T. P. (2010) Fundamentals of Digital Image Processing: a Practical Approach with Examples in Matlab. USA, Wiley-Blackwell Publ. DOI: 10.1002/9780470689776.
5. Gonzalez R. (2018) Digital Image Processing. New York, Pearson Publ.
6. Ausiannikau A. V., Kozel V. M. (2021) Filtration of Histogram Evaluation of Probability Density Based on Fuzzy Data Accessibility to a Grouping Interval. Doklady BGUIR. 19 (4), 13–20. https://doi.org/10.35596/17297648-2021-19-4-13-20 (in Russian).
7. Ausiannikau A. V. (2009) Statisticheskie Neravenstva v Sverh Regulyarnyh Statisticheskih Eksperimentah Teorii Otsenivaniya. Vestsi Natsiyanal’naj Akademii Navuk Belarusi. Ser. Fiz-Mat. Navuk. (2), 106–110 (in Russian).
8. Novitskiy P. V., Zograf I. A. (1991) Evaluation of Errors of Measurement Results, 2-e izd. pererab. i dop. Leningrad, Energoatomizdat Publ., Leningrad. Otdelenie. 304 (in Russian).
Review
For citations:
Ausiannikau A.V., Kozel V.M. Histogram Filter with Smoothing Parameter Setting. Doklady BGUIR. 2022;20(8):42-50. (In Russ.) https://doi.org/10.35596/1729-7648-2022-20-8-42-50
Views: 341
Email this article 