Refine
Year of publication
- 2019 (1)
Document Type
- ZIB-Report (1)
Language
- English (1)
Has Fulltext
- yes (1)
Is part of the Bibliography
- no (1)
Institute
- Mathematical Optimization (1) (remove)
We propose a partially functional autoregressive model with exogenous variables
(pFAR) to describe the dynamic evolution of the serially correlated functional data.
It provides a united framework to model both the temporal dependence on multiple lagged functional covariates and the causal relation with ultrahigh-dimensional exogenous scalar covariates. Estimation is conducted under a two-layer sparsity assumption, where only a few groups and elements are supposed to be active, yet without knowing their number and location in advance. We establish asymptotic properties of the estimator and investigate its unite sample performance along with simulation studies. We demonstrate the application of pFAR with the high-resolution natural gas flows in Germany, where the pFAR model provides insightful interpretation as well as good out-of-sample forecast accuracy.