Hi,
I am trying to fit ARIMA models to a high number of time series', too many to do manually (at this stage). In an attempt to automate I am trying to pull out all the reasonable variations of models that each serise could be fit with. Part of this is extracting the specific lags for autocorrelation and partial autocorrelation estimates that land outside of the grey 95% confidence boundary as highlighted the pac and ac commands. The stored values of corrgram do not contain these standard errors or confidence intervals. Is there a way to extract or calculate these figures or another way to approximate the 95% confidence intervals?
An example:
webuse air2
corrgram air, lags(20)
ac d.air
pac d.air
So the figures I want to extra are the estimate grey areas for each lag in ac and pac commands.
Thanks very much in advance
Henry
I am trying to fit ARIMA models to a high number of time series', too many to do manually (at this stage). In an attempt to automate I am trying to pull out all the reasonable variations of models that each serise could be fit with. Part of this is extracting the specific lags for autocorrelation and partial autocorrelation estimates that land outside of the grey 95% confidence boundary as highlighted the pac and ac commands. The stored values of corrgram do not contain these standard errors or confidence intervals. Is there a way to extract or calculate these figures or another way to approximate the 95% confidence intervals?
An example:
webuse air2
corrgram air, lags(20)
ac d.air
pac d.air
So the figures I want to extra are the estimate grey areas for each lag in ac and pac commands.
Thanks very much in advance
Henry
