IDENTIFYING MULTIVARIATE TIME SERIES MODELS
Authors:
- 1 May 1982
- journal article
- Published by Wiley in Journal of Time Series Analysis
- Vol. 3 (3) , 153-164
Abstract
Akaike (1974, 1975) has described how canonical variate analysis can be used to identify the structure of linear multivariate time series models. With some modification, the procedure is suitable for finding autoregressive moving average representations which are efficiently parameterized. We describe briefly the method and examine its performance when applied to a well‐known bivariate time series.Keywords
This publication has 8 references indexed in Scilit:
- SOME ASPECTS OF MODELLING AND FORECASTING MULTIVARIATE TIME SERIESJournal of Time Series Analysis, 1981
- The Multivariate Portmanteau StatisticJournal of the American Statistical Association, 1980
- The Multivariate Portmanteau StatisticJournal of the American Statistical Association, 1980
- Multiple Time Series Modelling: Another Look at the Mink-Muskrat InteractionJournal of the Royal Statistical Society Series C: Applied Statistics, 1978
- Canonical Correlation Analysis of Time Series and the Use of an Information Criterion1976
- Markovian Representation of Stochastic Processes by Canonical VariablesSIAM Journal on Control, 1975
- Markovian representation of stochastic processes and its application to the analysis of autoregressive moving average processesAnnals of the Institute of Statistical Mathematics, 1974
- Block Toeplitz Matrix InversionSIAM Journal on Applied Mathematics, 1973