Stationarity and Stability of Autoregressive Neural Network Processes

Abstract

We analyze the asymptotic behavior of autoregressive neural net(cid:173) work (AR-NN) processes using techniques from Markov chains and non-linear time series analysis. It is shown that standard AR-NNs without shortcut connections are asymptotically stationary. If lin(cid:173) ear shortcut connections are allowed, only the shortcut weights determine whether the overall system is stationary, hence standard conditions for linear AR processes can be used.

Cite

Text

Leisch et al. "Stationarity and Stability of Autoregressive Neural Network Processes." Neural Information Processing Systems, 1998.

Markdown

[Leisch et al. "Stationarity and Stability of Autoregressive Neural Network Processes." Neural Information Processing Systems, 1998.](https://mlanthology.org/neurips/1998/leisch1998neurips-stationarity/)

BibTeX

@inproceedings{leisch1998neurips-stationarity,
  title     = {{Stationarity and Stability of Autoregressive Neural Network Processes}},
  author    = {Leisch, Friedrich and Trapletti, Adrian and Hornik, Kurt},
  booktitle = {Neural Information Processing Systems},
  year      = {1998},
  pages     = {267-273},
  url       = {https://mlanthology.org/neurips/1998/leisch1998neurips-stationarity/}
}