Generalized entropy optimization distributions dependent on parameter in time series


Shamilov A., Giriftinoglu C.

WSEAS Transactions on Information Science and Applications, vol.7, no.1, pp.102-111, 2010 (Scopus) identifier

  • Publication Type: Article / Article
  • Volume: 7 Issue: 1
  • Publication Date: 2010
  • Journal Name: WSEAS Transactions on Information Science and Applications
  • Journal Indexes: Scopus
  • Page Numbers: pp.102-111
  • Keywords: Autoregressive modelling, Estimation of missing value, MaxEnt distribution, MinMaxEnt distribution, Time series with missing value
  • Anadolu University Affiliated: Yes

Abstract

In this paper, we have proposed Generalized Entropy Optimization Problems (GEOP) concerned with parameters which consist of maximizing Entropy Functional subject to constraints dependent on parameters. Therefore, MinMaxEnt and MaxMaxEnt distributions are obtained as solutions of these problems. On bases of MinMaxEnt distribution we have developed a new estimation method of obtaining missing values in time series. Mentioned estimation method is applied to the data generated from autoregressive model for one missing value. The performance of the developed method is evaluated by mean square errors (MSE) calculated from simulation studies and the validity of this method is shown.