Problem with out-of-sample fitted values from equation

[adrian_d]

Hi,
I encountered a problem with getting right fitted values variable from equations estimated on a part of sample. If I have data e.q. from Jan2014 to Oct2014, but estimate on May-Oct2014, I just want to create fitted variable with "NA" from Jan-April, but it seems impossible :( How to get this using command line?

Btw second question, is it possible to get "adjusted sample" from equation, if estimated sample is different than declared in equation specification (due to for example NA in one variables) ?

thanks
adrian

[EViews Gareth]

Not entirely sure I follow the first question, but did you try the (f=na) option on the fit command?

There isn't really a good way to get the adjusted sample, other than by making the residual series (which only has values for observations that were used in estimation - any observation with an NA is one that was not used).

[adrian_d]

Yup, I have tried (f=na) option on the fit command, it doesn't work. But your second idea is great, making resids is good enough for me ! Thanks :)