• Multivariate time series;
  • covariance estimation;
  • adaptive estimation;
  • dynamic linear models;
  • multivariate control charts

This article develops on-line inference for the multivariate local level model, with the focus being placed on covariance estimation of the innovations. We assess the application of the inverse Wishart prior distribution in this context and find it too restrictive since the serial correlation structure of the observation and state innovations are forced to be the same. We generalize the inverse Wishart distribution to allow for a more convenient correlation structure, but still retaining approximate conjugacy. We prove some relevant results for the new distribution and we develop approximate Bayesian inference, which allows simultaneous forecasting of time series data and estimation of the covariance of the innovations of the model. We provide results on the steady state of the level of the time series, which are deployed to achieve computational savings. Using Monte Carlo experiments, we compare the proposed methodology with existing estimation procedures. An example with real data consisting of production data from an industrial process is given.