A nonlinear model for time series prediction and signal interpolation

Author

Niranjan, Mahesan ; Kadirkamanathan, Visakan

Author_Institution

Dept. of Eng., Cambridge Univ., UK

fYear

1991

fDate

14-17 Apr 1991

Firstpage

1713

Abstract

The approach is an extension of the method of radial basis functions. Parameter estimation for the nonlinear predictor is performed by a gradient descent over a mean squared error measure, starting from a random initialization of the parameters. Results on predicting segments of speech data and the sunspot series are presented and compared to a linear predictor. An approach to adaptive estimation of the model by means of an extended Kalman filter is presented. In terms of prediction residual, the nonlinear predictor is found to perform significantly better than a linear model with the same number of parameters. Difficulties in applying this model in speech processing are discussed

Keywords

Kalman filters; filtering and prediction theory; interpolation; nonlinear systems; speech analysis and processing; time series; adaptive estimation; extended Kalman filter; gradient descent; linear predictor; mean squared error measure; nonlinear model; nonlinear predictor; prediction residual; radial basis functions; random initialization; signal interpolation; speech data; sunspot series; time series prediction; Adaptive estimation; Context modeling; Interpolation; Least squares approximation; Neural networks; Parameter estimation; Performance evaluation; Predictive models; Speech processing; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1991. ICASSP-91., 1991 International Conference on

Conference_Location

Toronto, Ont.

ISSN

1520-6149

Print_ISBN

0-7803-0003-3

Type

conf

DOI

10.1109/ICASSP.1991.150640

Filename

150640