Modeling and Forecasting of Multivariate Time Series
Gong Zhengf · Information and Computation · 1987
For an equivalent canonical form of multivariate time series transfer function model,a modeling and forecasting scheme,which consists of preliminary modeling and ExtendedKalman Adaptive Estimation-Prediction (EKAEP) algorithm,is presented in this paper.In consideration of time-variation and outliers of physical systems,both the noise stat-istic property Limited Memo(?)y Adaptive Estimator (IMAE) and the outlier detectionmethod based on the criterion of doube believable intervals are included in the EKAEPalgorithm.The efficiency of the scheme has been proved by several simulations.