[Octave-bug-tracker] [bug #62161] [octave forge] (tsa) acorf depends on constant

[Alois Schlögl]

Follow-up Comment #1, bug #62161 (project octave):



When you refer to a "constant", I guess you refer to non-zero mean of the time
series. For your example, any of the following would do  

   acorf(detrend(([0:10] + x),0)', 1)
   acorf(center ( (1:10) + x ,1)', 1)

The help function is quite verbose about how a non-zero mean could be handled.
(see 'help acorf') 

Therefore, I do not agree with your assessment, that the function returns an
incorrect result. The function does also not even have the same name than the
name a in matlab. So there also no confusion. 

If you believe there is a need for an 'autocorr' function, that shows the
behavior you expect, it would be very simple to add this. 










    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?62161>

_______________________________________________
  Message sent via Savannah
  https://savannah.gnu.org/