Forum: help

RE: How to calculate Durbin-Watson under systemfit? [ Reply ] By: Arne Henningsen on 2014-09-14 06:51	[forum:41451]
Dear Yu It is not difficult to calculate DW values but I am currently busy with other projects, Please do not hesitate to file a feature request or implement this feature in systemfit. Best regards, Arne

RE: How to calculate Durbin-Watson under systemfit? [ Reply ] By: Yu Zhang on 2014-08-18 00:08	[forum:41379]
Dear Arne, Thank you for your kind reply. I will appreciate it very much if you could tell me why it is difficult to calculate DW values under systemfit. I noticed that this issue had been asked several years ago. Thank again. Yu

RE: How to calculate Durbin-Watson under systemfit? [ Reply ]
By: Arne Henningsen on 2014-08-14 06:26

[forum:41372]

Dear Yu

Unfortunately, we haven't yet implemented the calculation of the Durbin-Watson statistic in systemfit. Hence, you have to do it manually. Sorry!

You could create a feature request so that this does not disappear from my to-do list:

https://r-forge.r-project.org/tracker/?atid=986&group_id=242

I agree that an implementation of the Parks (1967) estimator could be a useful enhancement of the systemfit package. However, I doubt that I will have sufficient time for this in the near future. You can also create a feature request for the Park estimator.

...and of course your are invited to join the systemfit team and to implement these features yourself.

Best wishes,
Arne

How to calculate Durbin-Watson under systemfit? [ Reply ]
By: Yu Zhang on 2014-08-12 04:39

[forum:41344]

Dear Arne,

I am using systemfit to estimate equations of supply and demand of a product in a region by using time series data, including quantity and price.
Now I wonder how to calculate Durbin-Watson value for each equation.

<The package could be enhanced by the inclusion of methods to t equations with serially correlated and heteroscedastic disturbances (Parks 1967).> Progress is expected. Thanks a lot.