st: Doubts about ARIMA forecast

["Ángela Fajardo Moreno" <aafajardom@unal.edu.co>]

Hi dear users,

I have the next doubts about my programmation and I would really appreciate for you to help me because I have not found any answer in historical statalist.

1) The program does not understand the stdp order after predict, this is what appear after running my arima model (arima dlipc, ar(1/2) ma(1 12)):

predict pdlipc if t>tm(2013m3), xb dynamic(tm(2013m2))

predict s, stdp
(option xb assumed; predicted values)

2) How do I program the confidence interval (at 95% level) for the forecast of the variable?

3) How to make a conditional forecast? I mean, if I want that the ipc variable have a value of 0.03 in 2013m12, how should be the values in 2013m4 - 2013m11 out-of-sample?

4) I need to compare the forecast of different ARIMA models to choose which one explain better the predict out-of-sample of the variable, Which statistics let me analyse that? For example how could I programming the Theil's U?

Thanks for your help,
Angela

Enviado desde mi iPad
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/