Re: [Gretl-users] a simple question about arima command output

between -1/2 and 1/2 ? The question is removed: data and forecasts are cyclic with period of abs(1/frequency) observations Still, I do not know whether this pattern does not depend on quantity of terms in a model Oleh 8 листопада 2018, 13:59:09, від oleg_komashko(a)ukr.net: I'll try to simplify my question My guess is  frequency= +-angle/(2pi)    0<= angle<=pi frequency = +-(2pi-angle)/(2pi) pi<angle<2pi 1) am I right? 2) If so, is it correct: why frequencies are always between -1/2 and 1/2 ? Oleh 8 листопада 2018, 13:22:40, від "Sven Schreiber" <svetosch(a)gmx.net&gt;: Am 08.11.2018 um 10:46 schrieb oleg_komashko(a)ukr.net: The matter is in that it is a function of the angle I have failed to guess what the function it is. That's why I have asked. I'm not sure I understand the question. If the root is 0.5 + i*0.1, then the angle should be atan(0.1 / 0.5) = 0.19739556 (in radians). Then you still have to factor in that the ARMA "roots" that are given are probably not the ones inside the unit circle (for stability), but in fact the inverse roots outside the circle. So before taking the atan you'd have to invert the root. Is that what you mean? thanks, sven _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users