Title :
A reduced parameter bilinear time series model
Author :
Zhang, Yongqing ; Hagan, Martin T.
Author_Institution :
Mactronix, Richardson, TX, USA
fDate :
7/1/1994 12:00:00 AM
Abstract :
A new bilinear time series structure is proposed and is tested on three sample time series to demonstrate its effectiveness. It is found that the proposed bilinear model can represent both nonlinearity and multiperiodicity, and it therefore provides a useful model class for general applications. In addition, the proposed bilinear model uses fewer parameters than conventional bilinear models with the same structure
Keywords :
parameter estimation; time series; bilinear time series; estimation techniques; identification techniques; multiperiodicity; nonlinearity; reduced parameter model; Analysis of variance; Autoregressive processes; Convolution; Digital signal processing; Gaussian noise; Integrated circuit modeling; Signal analysis; Signal processing; Telecommunications; White noise;
Journal_Title :
Signal Processing, IEEE Transactions on
