Neural network integration of classical statistical tests for processing small samples of biometrics data
https://doi.org/10.21683/1729-2646-2019-19-2-22-27
Abstract
The Aim of this paper is to increase the power of statistical tests through their joint application to reduce the requirement for the size of the test sample.
Methods. It is proposed to combine classical statistical tests, i.e. chi square, Cram r-von Mises and Shapiro-Wilk by means of using equivalent artificial neurons. Each neuron compares the input statistics with a precomputed threshold and has two output states. That allows obtaining three bits of binary output code of a network of three artificial neurons.
Results. It is shown that each of such criteria on small samples of biometric data produces high values of errors of the first and second kind in the process of normality hypothesis testing. Neural network integration of three tests under consideration enables a significant reduction of the probabilities of errors of the first and second kind. The paper sets forth the results of neural network integration of pairs, as well as triples of statistical tests under consideration.
Conclusions. Expected probabilities of errors of the first and second kind are predicted for neural network integrations of 10 and 30 classical statistical tests for small samples that contain 21 tests. An important element of the prediction process is the symmetrization of the problem, when the probabilities of errors of the first and second kind are made identical and averaged out. Coefficient modules of pair correlation of output states are averaged out as well by means of artificial neuron adders. Only in this case the connection between the number of integrated tests and the expected probabilities of errors of the first and second kind becomes linear in logarithmic coordinates.
About the Authors
A. I. Ivanov
Penza Research and Design Electrical Engineering Institute
Russian Federation
Russian Federation
Alexander I. Ivanov, Doctor of Engineering, Associate Professor, Lead Researcher of Laboratory of Biometric and Neural Network Technologies
Penza
E. N. Kuprianov
Penza State University
Russian Federation
Russian Federation
Evgeny N. Kuprianov, post-graduate student, Department of Information Security Technology
Penza
S. V. Tureev
Research and Design Institute for Communications and Control Systems
Russian Federation
Russian Federation
Sergey V. Tureev, Head of Research and Technology Center
Moscow
References
1. Pokhabov Yu.P. Problems of dependability and possible solutions in the context of unique highly vital systems design. Dependability 2019;19(1):10-17.
2. Pokhabov Yu.P. Ensuring dependability of unique highly vital systems. Dependability 2017;17(3):17-23,
3. GOST R 52633.5-2011. Information protection. Information protection technology. The neural net biometry-code converter automatic training. Moscow: Standartinform; 2012 [in Russian].
4. Kobzar A.I. Prikladnaia matematicheskaia statistika: dlia inzhenerov i nauchnykh rabotnikov [Applied mathematical statistics: for engineers and researchers]. Moscow: FIZMATLIT; 2006 [in Russian].
5. Iazov Yu.K., editor, Volchikhin V.I., Ivanov A.I., Funtikov V.A., Nazarov I.G. Neyrosetevaia zashchita personalnykh biometricheskikh dannykh [Neural network protection of biometric data]. Moscow: Radiotekhnika; 2012 [in Russian].
6. Sukhoruchenkov B.I. Analiz maloy vyborki. Prikladnye statisticheskie metody [Small sample analysis. Applied statistical methods]. Moscow: Vuzovskaya kniga; 2010 [in Russian].
7. Doerffel K. Statistics in analytical chemistry. Moscow: Mir; 1994.
8. Dayev Zh.A., Nurushev E.T. Application of statistical criteria for improving the efficiency of risk assessment methods. Dependability 2018;2:42-45.
9. Akhmetov B.B., Ivanov A.I. Estimation of quality of a small sampling biometric data using a more efficient form of the chi-square test. Dependability 2016;16(2):43-48.
10. R 50.1.037-2002. Rekomendatsii po standartizatsii. Prikladnaia statistika. Pravila proverki soglasia opytnogo raspredelenia s teoreticheskim. Chast I. Kriterii tipa χ2 [Standardization recommendations. Applied statistics. Rules for compliance verification of experimental distribution with the theoretical one. Part I. Chi-square-type criteria]. Moscow: Gosstandart of Russia; 2001 [in Russian].
11. R 50.1.037-2002. Prikladnaia statistika. Pravila proverki soglasia opytnogo raspredelenia s teoreticheskim. Chast II. Neparametricheskie kriterii [Applied statistics. Rules for compliance verification of experimental distribution with the theoretical one. Part II. Non-parametric tests]. Moscow: Gosstandart of Russia; 2002 [in Russian].
12. Haykin S. Neural Networks: A Comprehensive Foundation. Moscow: Viliams; 2006.
13. Serikova N.I., Ivanov A.I., Serikova Yu.I. Otsenka pravdopodobia gipotezy o normalnom raspredelenii po kriteriu Dzhini dlia chisla stepeney svobody, kratnogo chislu opytov [Likelihood estimation of the hypothesis of normal Gini distribution for the number of degrees of freedom multiple of the number of experiments]. Voprosy radioelektroniki 2015;1(1):85-94 [in Russian].
14. Perfilov K.A. Kriteriy srednego geometricheskogo, ispolzuemyy dlia proverki dostovernosti statisticheskikh gipotez raspredelenia biometricheskikh dannykh [Criterion of geometric mean used for validity verification of the statistical hypotheses of biometric data distribution]. Proceedings of the science and technology conference of the Penza information technology security cluster. Penza; 2014. Vol. 9. p. 92-93, <http://www.pniei.penza.ru/RV-conf/T9/S92> [in Russian].
15. Ivanov A.I., Perfilov K.A. Otsenka sootnoshenia moshchnostey semeystva statisticheskikh kriteriev “srednego geometricheskogo” na malykh vyborkakh biometricheskikh dannykh [Assessment of the relations of the strengths of the “geometric mean” family of statistical tests on small samples of biometric data]. XI All-Russian Science and Practice Conference Modern security technologies and comprehensive facility security assets. Penza-Zarechny; 2016. p. 223-229 [in Russian].
16. Ivanov A.I., Perfilov K.A., Malygina E.A. Multivariate statistical analysis of the quality of biometric data on extremely small samples using the criteria of the geometric mean tests calculated for the analyzed probability functions. Measuring. Monitoring. Management. Control 2016;2(16):64-72 [in Russian].
17. Ivanov A.I., Perfilov K.A., Malygina E.A. Evaluation of the quality of small samples of biometric data using a differential variant statistical test of the geometric mean. Vestnik SibGAU 2016;4(17):864-871 [in Russian].
18. Malygin A.Yu., Volchikhin V.I., Ivanov A.I., Funtikov V.A. Bystrye algoritmy testirovania neyrosetevykh mekhanizmov biometriko-kriptograficheskoy zashchity informatsii [Fast algorithms for testing of neural network mechanisms of biometric and cryptographic protection of information]. Penza, Penza State University Publishing 2006 [in Russian].
19. Akhmetov B.S., Volchikhin V.I., Ivanov A.I., Malygin A.Yu. Algoritmy testirovania biometriko-neyrosetevykh mekhanizmov zashchity informatsii [Algorithms for testing biometric and neural network mechanisms of information protection]. Almaty: Satpayev KazNTU; 2013 [in Russian].
Review
For citations:
Ivanov A.I., Kuprianov E.N., Tureev S.V. Neural network integration of classical statistical tests for processing small samples of biometrics data. Dependability. 2019;19(2):22-27. https://doi.org/10.21683/1729-2646-2019-19-2-22-27
Views:
975
ISSN 1729-2646 (Print)
ISSN 2500-3909 (Online)
ISSN 2500-3909 (Online)
Email this article 