|
Abstract:
|
In economics, finance and many other scientific fields, observations at each time point may naturally take the form of a matrix, with potentially large dimensions. We propose an autoregressive model for the matrix-valued time series. It preserves the matrix structure and admits corresponding interpretations. Comparing with classical vector autoregressive models, it is also a more parsimonious model because of the matrix structures and the additional interpretable low rank or sparsity assumptions on the coefficient matrices. Estimation procedures and their theoretical properties are investigated. The performances of the model are demonstrated with simulated and real examples.
|
