Online algorithms for modeling distributions using examples

Author

Thathachar, M. A L ; Arvind, M.T.

Author_Institution

Dept. of Electr. Eng., Indian Inst. of Sci., Bangalore, India

Volume

3

fYear

1997

fDate

9-12 Sep 1997

Firstpage

1315

Abstract

This paper addresses the problem of modeling the relationships between observed samples of data as a distribution. An L-step dependent model is constructed and an online algorithm is designed for the model in order to minimize the Kullback measure. The algorithm is analyzed to show that it converges weakly to global optimum of Kullback measure for that model. Simulation studies indicate that the algorithm has better tracking properties for time-varying distributions, when compared with statistical estimation procedures

Keywords

convergence of numerical methods; signal sampling; statistical analysis; time series; Kullback measure; binary strings; convergence analysis; distribution modeling; global optimum; observed samples; online algorithms; simulation; statistical estimation; time series; time-varying distributions; tracking properties; Adaptive algorithm; Algorithm design and analysis; Convergence; Data engineering; Learning systems; Position measurement; Predictive models; Probability; Speech recognition; Stock markets;

fLanguage

English

Publisher

ieee

Conference_Titel

Information, Communications and Signal Processing, 1997. ICICS., Proceedings of 1997 International Conference on

Print_ISBN

0-7803-3676-3

Type

conf

DOI

10.1109/ICICS.1997.652201

Filename

652201