| The Statistical Software spectrum - version 3.3 © | ||||||||
| Partialization analysis and graphical models for stochastic processes | ||||||||
|
The statistical software spectrum - version 3.3 allows to investigate the linear interrelations between the components of a multivariate time series. The program computes and plots oridinary and partialized second-order statistics in the frequency domain (partial spectra and partial coherences) and the time domain (partial correlations). The partial coherences allow the identification of the partial correlation graph of the multivariate process (e.g. Dahlhaus 2000). The partial statistics can also be used to identify directed dependencies (Dahlhaus et al. 1997, Eichler et al. 2000). All graphics are formatted in PostScript and can be either displayed using e.g. ghostscript or printed by e.g. lpr. | ||||||||
| Papers on the subject | ||||||||
| ||||||||
| About the software | ||||||||
|
The software is implemented in C and requires an UNIX system.
However, it is written device independent since all graphical results are
formatted in PostScript and written in output files. For more details on the program
please refer to the manual page. | ||||||||
| Where to get the software | ||||||||
|
The software is distributed via our anonymous ftp-server. Please feel free to copy it. | ||||||||
| ||||||||
|
Any feedback on successful applications of the program is welcome. Author of the software: Michael Eichler | ||||||||
| ||||||||
|
| ||||||||
| ||||||||
| ||||||||