EViews.com

Hi Matt,

Last question, the code is working however I'm loosing some of the data from the start of the series.

For example, the threshold starts 01 mars 1974 however the signal is starting only from 14 october 1977.

I have enclosed the program and the expected series.

Thanks alots.

It's really appreciated to help me on your free time.

Sudesh

That's almost certainly due to the lag dependence in the signal, i.e., for the first observation in your sample {%sig}(-1) is NA, which propagates to later observations until you get to a case number that doesn't depend on {%sig}(-1), such as case number 9 in the example you posted. In the statement that generates the signal values, try replacing all the "{%sig}(-1)" with "@nan({%sig}(-1), 0)". That will cause NAs to be treated as a signal of 0.

Thanks alots Matt..