Wyniki wyszukiwania - BazTech

1

Forecasting GDP growth rate in Ukraine with alternative models

Karayuz I, Bidyuk P.

Applied Computer Science

|

2015

|

Vol. 11, no 3

88--97

EN

The problem of constructing mathematical model for short-term fore-casting of GDP is considered. First, extended autoregression is constru-cted that takes two additional independent variables into consideration. The model resulted provides a possibility for generating short-term forecasts of GDP though not of high quality. Another model was constructed in the form of a Bayesian network. The model turned out to be better than the multiple regression, it provides quite good estimates for probabilities of GDP growth direction.

2

Effective expectation maximization algorithm implementation using multicore computer systems

Kasitskij A., Bidyuk P., Gozhyi A.

Informatyka, Automatyka, Pomiary w Gospodarce i Ochronie Środowiska

|

2014

|

nr 4

35--37

EN

A popular expectation maximization algorithm that is widely used in modern data processing systems to solve various problems including optimization and parameter estimation is considered in the paper. The task of the study was to enhance effectiveness of the algorithm execution in time. An enhancement of execution rate for the EM algorithm using multicore architecture of modern computer systems was carried out. Necessary modifications aimed at better parallelism were proposed for implementation of the EM algorithm. An efficiency of the software implementation was tested on the classic problem of Gaussian random variables mixture separation. It is shown that in the mixture separation problem EM algorithm performance degrades when the distance between mean values of distributions is less than three standard deviations, which is totally in the spirit of three sigma law. In such cases, it is very important to have an efficient EM algorithm implementation to be able to process such test cases in a reasonable time.

PL

W artykule opisany jest popularny algorytm EM (expectation maximization), który jest powszechnie stosowany w nowoczesnych systemach przetwarzania danych do rozwiązywania różnych problemów, w tym optymalizacji i estymacji parametrów. Celem badań było zwiększenie efektywności czasu wykonywania algorytmu. Zwiększenie szybkości wykonania algorytmu EM użyto wielordzeniowy architektury nowoczesnych systemów komputerowych. Zostały zaproponowane niezbędne modyfikacje mające na celu lepszą równoległość realizacji algorytmu EM. Skuteczność implementacji programu była testowana na klasycznym problemie separacji Gaussowskich zmiennych losowych. Wykazano, że w przypadku rozdziału mieszaniny wydajność algorytmu EM ulega degradacji, kiedy odległość między średnimi wartościami rozkładu wynosi mniej niż trzy odchylenia standardowe, co jest całkowicie zgodnie z regułą trzech sigm. W takich przypadkach, jest bardzo ważne, aby mieć efektywną realizację algorytmu EM móc przetworzyć takie przypadki w rozsądnym czasie.

3

Solving Parabolic Equations by Using the Method of Fast Convergent Iterations

Bondarenko V., Bidyuk P., Bernatska J.

International Journal of Applied Mathematics and Computer Science

|

2000

|

Vol. 10, no 2

333-344

EN

The paper describes an approach to solving parabolic partial differential equations that generalizes the well-known parametrix method. The iteration technique proposed exhibits faster convergence than the classical parametrix approach. A solution is constructed on a manifold with the application of the Laplace-Beltrami operator. A theorem is formulated and proved to provide a basis for finding a unique solution. Simulation results illustrate the superiority of the proposed approach in comparison with the classical parametrix method.

4

A stochastic criterion for the optimal investment process

Bondarenko J., Bidyuk P.

Applied Mathematics and Computer Science

|

1998

|

Vol. 8, no 3

505-510

EN

The paper describes an approach to the construction of a criterion suitable for optimization of an investment project. An income growth resulting from reaIization of the investment project is considered as a time sequence of random variabIes. Then an expression is derived which characterizes the project cost at a sequence of time intervals it and raises a possibility to find the real income as a random variable. Two situations are considered, namely when the investor possesses complete and incomplete information about the economic situation and the corresponding incomes. The obtained results can be used to construct an investment criterion in the case of a vague economic situation.