INTELLIGENT SYSTEMS AND TECHNOLOGIES
Yu.S. Popkov Entropy methad of compression matrices with random elements
COMPUTING SYSTEMS
PATTERN RECOGNITION
CONTROL AND DECISION-MAKING
Yu.S. Popkov Entropy methad of compression matrices with random elements

Abstract

Two-stages method of compression matrices with random elements is proposed. Operation of direct and invers projections are the kernal of method. The information cross-entropy is used for optimization these operations. Serial and parallel procedures are proposed.

Keywords:

cross-entropy, projection operator, matrix derivatives, gradient method.

pp. 8-15

References

1. Aivazian S.A, Buchshtaber V.M., Eniukov I.S., Meshalkin L.D. Prikladnaia statistika: klassifikacia i snizhenie razmernosti. М.: Finansi I statistika, 1989.
2. Merkov A.B. Raspoznavanie obrazov. Vvedenie v metodi statisticheskogo obuchenia. M. Editorial URSS, 2010.
3. Imelbaev S.S., Shmulian B.L. Modelirovie stochasticheskih kommunikacionnih sistem. М. Nauka, 1978, с.170-234.
4. Bruckstein A.M., Donoho D.L., Elad M. From Sparse Solutions of Systems of Equations to Sparse Modeling of Signals and Images. SIAM Rev. 2009, v.51, No.1, p.34-81.
5. Kendal M.J., Stewart A. Statisticheskie vivody I sviazi. М. Nauka, 1973.
6. Jollife I.T. Principal Component Analysis. N.Y. Springer-Verlag, 2002.
7. Polyak B.T., Hlebnikov M.V. Metod glavnih komponent: robastnie versii // Avtomatika I telemehanika, 2017, №3, с.130-148.
8. Vorontsov K.V. Matematicheskie metodi obuchenia po precedentam. Kurs lekcii MFTI, 2006.
9. Jaynes E.T. Information Theory and Statistical Mechanics. Physics Review, 1957, v.106, p. 620-639.
10. Popkov Yu.S. Macrosystem Theory and Applications. Springer, Lecture Notes in Informatics and Control, 1995.
11. Kullback S., Leibler R.A. On information and Sufficiency. Ann. of Math. Statistics, 1951, v.22(1), p. 79-86.
12. Magnus J., Neudecker H. Matrix Differential Calculus with Applications in Statistics and Econometrics. John Willey and Sons, 2007, thied Edt.

2019 / 03
2019 / 02
2019 / 01
2018 / 04

© ФИЦ ИУ РАН 2008-2018. Создание сайта "РосИнтернет технологии".