Hi Leon, if you send me your data I will estimate all the spectral regression fractions d at any ordinate of your choice using my own programs. Remember that the Geweke Porter-Hudak confidence limits (which I can also provide you with) are not correct as previously thought. I can also give you the true confidence limits, but suggest you read my vol 31 number 9 yr2002 Communications in Statistics article upon bias etc, and the numerical integration results I reported in the article. I hope that might be a help to you. Richard Hudson Qld Australia ----- Original Message ----- From: "Liang Ding" <[log in to unmask]> To: <[log in to unmask]> Sent: Wednesday, March 16, 2005 7:17 AM Subject: questions about ARFIMA > Hi: > > I was using ARFIMA packages to estimate d and found some questions. > > 1, if I change starting value of d to some number greater than 0.5, the > program does not work. and it seems that the package cannot estimate d > that > would be greater than 0.5. how to solve this problem? > > 2. when I tried to use estimateGPH command to estimate d, what is the > constant > number for fPrint (the third number needed in the function)? > > I will appreciate it very much if someone can help me with those > questions. > > Liang (Leon) Ding > Department of Economics > Univeristy of North Carolina > at Chapel Hill > 919-914-6851 > > > -- > No virus found in this incoming message. > Checked by AVG Anti-Virus. > Version: 7.0.308 / Virus Database: 266.7.2 - Release Date: 11/03/2005 > > -- No virus found in this outgoing message. Checked by AVG Anti-Virus. Version: 7.0.308 / Virus Database: 266.7.2 - Release Date: 11/03/2005