Hello, When I run the garch examples, I get the following output:
> dax.garch <- garch(dax) ***** ESTIMATION WITH ANALYTICAL GRADIENT ***** /// TUNER5.... V(30) = 0.750E+00 SHOULD BE BETWEEN 0.503-314 AND NaN /// AFCTOL.... V(31) = 0.100E-19 SHOULD BE BETWEEN 0.518-317 AND NaN /// LMAX0..... V(35) = 0.100E+01 SHOULD BE BETWEEN 0.518-317 AND NaN /// LMAXS..... V(36) = 0.100E+01 SHOULD BE BETWEEN 0.518-317 AND NaN /// DINIT..... V(38) = -0.100E+01 SHOULD BE BETWEEN -0.100E+02 AND NaN /// DTINIT.... V(39) = 0.100E-05 SHOULD BE BETWEEN 0.000E+00 AND NaN /// D0INIT.... V(40) = 0.100E+01 SHOULD BE BETWEEN 0.000E+00 AND NaN Warning message: `Machine' is deprecated. Use `.Machine' instead. See ?Deprecated. > I canšt figure out what to do about the machine is deprecated issue. It seems that the routine is not picking up machine numerical characteristics, eps etc. I tried an eps = 0.1 and get slightly better errors (see below). So, when it calls dsumsl, the errors above are produced. I have checked str(.Machine) and the definitions are all there. What should I do to get garch to run? Regards, Nick With eps = 0.1 ***** ESTIMATION WITH ANALYTICAL GRADIENT ***** /// TUNER5.... V(30) = 0.750E+00 SHOULD BE BETWEEN 0.503-314 AND NaN /// AFCTOL.... V(31) = 0.100E-01 SHOULD BE BETWEEN 0.518-317 AND NaN /// RFCTOL.... V(32) = 0.215E+00 SHOULD BE BETWEEN 0.503-314 AND 0.100E+00 /// XFTOL..... V(34) = 0.100E+02 SHOULD BE BETWEEN 0.000E+00 AND 0.100E+01 /// LMAX0..... V(35) = 0.100E+01 SHOULD BE BETWEEN 0.518-317 AND NaN /// LMAXS..... V(36) = 0.100E+01 SHOULD BE BETWEEN 0.518-317 AND NaN /// DINIT..... V(38) = -0.100E+01 SHOULD BE BETWEEN -0.100E+02 AND NaN /// DTINIT.... V(39) = 0.100E-05 SHOULD BE BETWEEN 0.000E+00 AND NaN /// D0INIT.... V(40) = 0.100E+01 SHOULD BE BETWEEN 0.000E+00 AND NaN -- Nicholas and Silvia Waltner 1726 Evergreen Place Seattle, WA 98122 Cellular: 206.972.1456 Phone/Fax: 206.726.3486 Email: [EMAIL PROTECTED] [[alternate HTML version deleted]] ______________________________________________ [EMAIL PROTECTED] mailing list http://www.stat.math.ethz.ch/mailman/listinfo/r-help