Transcription of Sparse Vector Autoregressive Modeling
{{id}} {{{paragraph}}}
Example: biology
Sparse Vector Autoregressive Modeling
Sparse Vector Autoregressive ModelingRichard A. Davis, Pengfei Zang, Tian ZhengDepartment of statistics, Columbia UniversityJuly 1, 2012 AbstractThe Vector Autoregressive (VAR) model has been widely used for Modeling temporal de-pendence in a multivariate time series. For large (and even moderate) dimensions, the numberof AR coefficients can be prohibitively large, resulting in noisy estimates, unstable predictionsand difficult-to-interpret temporal dependence. To overcome such drawbacks, we propose a2-stage approach for fitting Sparse VAR (sVAR) models in which many of the AR coefficientsare zero. The first stage selects non-zero AR coefficients based on an estimate of the partialspectral coherence (PSC) together with the use of BIC. The PSC is useful for quantifying theconditional relationship between marginal series in a multivariate process. A refinement secondstage is then applied to further reduce the number of parameters.
Sparse Vector Autoregressive Modeling Richard A. Davis, Pengfei Zang, Tian Zheng Department of statistics, Columbia University July 1, 2012 Abstract
Loading..
Tags:
Information
Domain:
Source:
Link to this page:
Please notify us if you found a problem with this document: