Abstract:
We propose a new randomized projection method based on entropy optimization of random projection matrices (the ERP method). The concept of compactness indicator of a data matrix, which is stored in projection matrices, is introduced. An ERP algorithm is formulated in the form of a conditional maximization problem for an entropy functional defined on the probability density functions of the projection matrices. A discrete version of this problem is considered, and conditions are obtained for the existence and uniqueness of its positive solution. Procedures are developed for the implementation of entropy-optimal projection matrices by sampling the probability density functions.
Keywords:
random projection, compression and expansion of data matrix, projection matrix, compactness indicator, density function sampling, variance set, interquartile set.
This publication is cited in the following 2 articles:
A. V. Kvasnov, A. A. Baranenko, E. Yu. Butyrskii, U. P. Zaranik, “O vliyanii tsentralnoi tendentsii na kharakter plotnosti raspredeleniya maksimalnoi entropii v mashinnom obuchenii”, Vestn. S.-Peterburg. un-ta. Ser. 10. Prikl. matem. Inform. Prots. upr., 19:2 (2023), 176–184
Xiao Xu, Hao Yan, Xiaolei Wang, Punit Gupta, “Research on Postproduction of Film and Television Based on Computer Multimedia Technology”, Scientific Programming, 2022 (2022), 1