Talks and Poster Presentations (without Proceedings-Entry):
Ch. Flamm, A. Graef, M. Deistler:
"Influence analysis for high-dimensional time series based on Granger-causality analysis";
Talk: 1st Austrian Stochastic Days,
Linz;
2012-09-24
- 2012-09-25.
English abstract:
Granger causality is a useful concept for studying causal relations in multivariate time series.
The corresponding methodology is based on regular autoregressive systems. However, prob-
lems occur when these systems are singular, i.e. the covariance matrix of the innovations is
1
Created from the Publication Database of the Vienna University of Technology.