From: CynthiaMcN@aol.com To: "RATS Discussion List" Subject: times series scramble Date: Sat, 7 Sep 1996 10:25:24 -0400 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: X-Mailer: Mercury MTS (Bindery) v1.30 I would like to "scramble" a data set: single series, 2500 observations. How would I do this with RATS? Thanks ---------- End of message ---------- From: Daniel McMillen To: "RATS Discussion List" Subject: Re: times series scramble Date: Sat, 7 Sep 1996 16:37:54 -0500 (CDT) Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: X-Mailer: ELM [version 2.4 PL25 PGP3 TCS] (via Mercury MTS (Bindery) v1.30) MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit There's probably a better way to scramble data, but try this: SET X = %RAN(1.) SET Y = T ORDER X / Y SET NEWVAR = VAR(FIX(Y)) This replaces VAR with a randomly drawn ordering, stored in NEWVAR. > Dan McMillen Tulane University > I would like to "scramble" a data set: single series, 2500 observations. > How would I do this with RATS? Thanks > ---------- End of message ---------- From: lmahens@vnet3.vub.ac.be (Luc M.A. Hens) To: "RATS Discussion List" Subject: Estima homepage Date: Mon, 9 Sep 1996 12:05:42 +0200 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Mailer: Mercury MTS (Bindery) v1.30 Hi I'm still trying to access the Estima homepage at http://www.estima.com. For weeks now, I have been getting the message: "server does not have a DNS entry." Other locations in Europe and the US are readily accessible from my computer. The problem also occurs when I try to access the Estima homepage from other servers in Belgium. According to Tom Maycock from Estima the problem may lie within 'InterNic' which is the DNS service for the United States and Canada. Do other Europeans on this list have similar problems? Could you let me (and the people at Estima) know, so that something can be done about it? Thanks, Luc Hens =================================================== Luc M.A. Hens assistant professor of economics Vesalius College, Vrije Universiteit Brussel Pleinlaan 2, B-1050 Brussels Belgium phone +32-2-629 20 61 fax +32-2-629 20 60 e-mail lmahens@vnet3.vub.ac.be My homepage : http://bedr.vub.ac.be/Luc_Hens.html VUB's homepage : http://www.vub.ac.be/VUB-home.html =================================================== ---------- End of message ---------- From: "F.Y.Kumah" To: "RATS Discussion List" Subject: Re: Estima homepage Date: Mon, 9 Sep 1996 12:14:04 MET Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Organization: Tilburg University X-mailer: Pegasus Mail for Windows (v2.23) (via Mercury MTS (Bindery) v1.30) Hello Luc, I have just logged in to Estima from The Netherlands. I have not had the error message you mentioned in your mail.. Regards Francis > Hi > > I'm still trying to access the Estima homepage at http://www.estima.com. > For weeks now, I have been getting the message: "server does not have a DNS > entry." Other locations in Europe and the US are readily accessible from > my computer. The problem also occurs when I try to access the Estima > homepage from other servers in Belgium. According to Tom Maycock from > Estima the problem may lie within 'InterNic' which is the DNS service for > the United States and Canada. > Do other Europeans on this list have similar problems? Could you let me > (and the people at Estima) know, so that something can be done about it? > > Thanks, > > Luc Hens > > =================================================== > Luc M.A. Hens > assistant professor of economics > Vesalius College, Vrije Universiteit Brussel > Pleinlaan 2, B-1050 Brussels Belgium > phone +32-2-629 20 61 fax +32-2-629 20 60 > e-mail lmahens@vnet3.vub.ac.be > My homepage : http://bedr.vub.ac.be/Luc_Hens.html > VUB's homepage : http://www.vub.ac.be/VUB-home.html > =================================================== > > > **~~**~~**~~**~~**~~**~~**~~**~~**~~****~~**~~**~~**~~** Francis Y. Kumah CentER for Economic Research P. O. Box 90153 5000 LE Tilburg The Netherlands Phone (+31) 13 - 4668 221 Fax (+31) 13 - 4663 066 My Home Page: http://cwis.kub.nl/~few5/center/phd_stud/kumah/home.HTM **~~**~~**~~**~~**~~**~~**~~**~~**~~****~~**~~**~~**~~** ---------- End of message ---------- From: Sune Karlsson To: "RATS Discussion List" Subject: Re: Estima homepage Date: Mon, 09 Sep 1996 12:45:12 +0100 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Organization: Stockholm School of Economics MIME-version: 1.0 X-Mailer: Mozilla 3.0 (WinNT; I) (via Mercury MTS (Bindery) v1.30) Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7bit Luc M.A. Hens wrote: > > Hi > > I'm still trying to access the Estima homepage at http://www.estima.com. > For weeks now, I have been getting the message: "server does not have a DNS > entry." Other locations in Europe and the US are readily accessible from > my computer. The problem also occurs when I try to access the Estima > homepage from other servers in Belgium. According to Tom Maycock from > Estima the problem may lie within 'InterNic' which is the DNS service for > the United States and Canada. > Do other Europeans on this list have similar problems? Could you let me > (and the people at Estima) know, so that something can be done about it? No problems here (Sweden), a DNS lookup shows the IP-number for www.estima.com as 204.148.147.94. You could always get there by using http://204.148.147.94/. Also, the authoritative name servers for estima.com are ns1.interaccess.com and ns2.interaccess.com, with ip-numbers 198.80.0.6 and 198.80.0.11. My guess is that your local DNS-server doesn't have www.estima.com in it's cache. When you try to access the page the DNS-server has to go out and ask a bunch of other servers for the ip-number, which takes some time. Your browser/ip-stack/whatever then probably times out before the server has finished the lookup process. If this is the problem, the server *should* eventually get the information and keep it in it's cache for a while. In other words, trying again after a few minutes *should* work. If not, DNS is probably misconfigured - the only question is where? /Sune -- Sune Karlsson | stsk@hhs.se Fax: + 46 8 34 81 61 Stockholm School of Economics | Phone: + 46 8 736 92 39 Box 6501, 113 83 Stockholm, Sweden | http://www.hhs.se/stat/ ---------- End of message ---------- From: CynthiaMcN@aol.com To: "RATS Discussion List" Subject: times series scramble Date: Sat, 7 Sep 1996 10:25:24 -0400 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: X-Mailer: Mercury MTS (Bindery) v1.30 (via Mercury MTS (Bindery) v1.30) I would like to "scramble" a data set: single series, 2500 observations. How would I do this with RATS? Thanks ---------- End of message ---------- From: Daniel McMillen To: "RATS Discussion List" Subject: Re: times series scramble Date: Sat, 7 Sep 1996 16:37:54 -0500 (CDT) Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: X-Mailer: ELM [version 2.4 PL25 PGP3 TCS] (via Mercury MTS (Bindery) v1.30) (via Mercury MTS (Bindery) v1.30) MIME-Version: 1.0 Content-Transfer-Encoding: 7bit There's probably a better way to scramble data, but try this: SET X = %RAN(1.) SET Y = T ORDER X / Y SET NEWVAR = VAR(FIX(Y)) This replaces VAR with a randomly drawn ordering, stored in NEWVAR. > Dan McMillen Tulane University > I would like to "scramble" a data set: single series, 2500 observations. > How would I do this with RATS? Thanks > ---------- End of message ---------- From: "Estima" To: "RATS Discussion List" Subject: Re: times series scramble Date: Mon, 9 Sep 1996 14:45:51 -0600 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT X-mailer: Pegasus Mail for Windows (v2.31) (via Mercury MTS (Bindery) v1.30) Regarding the data scrambling, Dan's suggestion is as good as any I know of: > There's probably a better way to scramble data, but try this: > SET X = %RAN(1.) > SET Y = T > ORDER X / Y > SET NEWVAR = VAR(FIX(Y)) > However, you can save a step or two by simply sorting your variable along with the random series. For example, if VAR is the series you want to scramble, just do: SET X = %RAN(1.0) ORDER X / VAR Note that if you wanted to draw randomly from your series _with replacement_ (as opposed to simply shuffling the given data), you should use the BOOT instruction instead. Thanks, Tom Maycock Estima -- +-----------------------------+-----------------------------------------+ | Estima | | | P.O. Box 1818 | Voice: (847) 864-8772 | | Evanston, IL 60204-1818 | Fax: (847) 864-6221 | | U.S.A | BBS: (847) 864-8816 | | e-mail: estima@estima.com | CompuServe: 73140,2202 | |-----------------------------------------------------------------------| | Web Site: http://www.estima.com | | RATS Internet Mailing List: New members can join by sending e-mail to | | MAISER@EFS.MQ.EDU.AU with the message: SUBSCRIBE RATS-L | +-----------------------------------------------------------------------+ ---------- End of message ---------- From: Economic Group RBA To: "RATS Discussion List" Subject: RATS for Macintosh Date: Thu, 12 Sep 1996 07:59:33 +1000 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Mailer: Mercury MTS (Bindery) v1.30 Any news yet on the long overdue power mac native version? Economic Analysis Department Email enquiries: Economic Group PH: +612 9551 8817 Reserve Bank of Australia FAX: +612 9551 8000 ---------- End of message ---------- From: Rob Trevor To: "RATS Discussion List" Subject: Re: RATS for Macintosh Date: Thu, 12 Sep 1996 14:07:57 +1000 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Mailer: Mercury MTS (Bindery) v1.30 At 7:59 AM 12/9/96, Economic Group RBA wrote: >Any news yet on the long overdue power mac native version? ^^^^^ Don't want to spread gloom, but its not as long overdue as ANY Mac version of something more recent than 4.02 (mine is dated Jan 93)! (An initial beta of 4.10 was released around May 94, but I've seen nothing more - despite some hopeful noises - in the last 2.3 years! I'm also unaware of any Mac version of RatsData which supports v4 databases.) Tom's not going to like this - but if you want to see Estima support the Mac, I'd suggest you yell *loudly*! Estima is a business, they need to make commercial decisions to support their actual and potential customer base. You need to show them that the Mac side of their business is important. Otherwise it doesn't make sense to support it. I run RATS on RS6000 workstations, Pentiums and Macs. The main reason I bought the Pentiums last December (rather than more Macs) is that the Mac version of RATS hasn't been updated for a long time. It was easier for me to switch hardware than to switch software (since I need to be able to run the same thing on the workstations). However, given the much higher cost of supporting (and the lower productivity of) Pentium hardware than Mac hardware on a TCP/IP network it was a tough decision. Next time I might look harder at the software alternatives. (Developing RATS type stuff under Windows with its available tools is a poor alternative to a Mac with tools like BBedit.) Cheers Rob PS Tom: I hope your mailbox fills up fast with responses from Mac users :) It will provide more paper with which to cover Tom D's desk. ----------------------------------------------------------------------- Rob Trevor Associate Professor Centre for Studies in Money, Banking and Finance Macquarie University, NSW, AUSTRALIA 2109 Internet: robt@efs.mq.edu.au VOICE: +61 (2) 9850-8447 FAX: +61 (2) 9850-7281 or +61 (2) 9418-1847 ----------------------------------------------------------------------- ---------- End of message ---------- From: Norman Morin To: "RATS Discussion List" Subject: Re: RATS for Macintosh Date: Wed, 11 Sep 1996 22:38:04 -0700 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Mailer: Mercury MTS (Bindery) v1.30 At 2:07 PM +1000 on 9/12/96, Rob Trevor wrote: >I run RATS on RS6000 workstations, Pentiums and Macs. The main reason I >bought the Pentiums last December (rather than more Macs) is that the Mac >version of RATS hasn't been updated for a long timeÉ(Developing RATS type >stuff under Windows with its available tools is a poor alternative to a Mac >with tools like BBedit.) Rob, I couldn't agree more; I use RATS on UNIX, PCs, and Macs, and am so much more productive on a Mac (w/ BBedit) it is barely comparable. I would love to see a PPC native version of RATS. --Norm Norman J. Morin nmorin@weber.ucsd.edu ------------------------------------------------------------- Department of Economics * University of California, San Diego 9500 Gilman Drive * La Jolla, CA 92023-0508 ------------------------------------------------------------- ---------- End of message ---------- From: Norman Morin To: "RATS Discussion List" Subject: Fractional Integration Date: Wed, 11 Sep 1996 22:58:01 -0700 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Mailer: Mercury MTS (Bindery) v1.30 Has anyone yet programmed Sowell's maximum likelihood estimation of a fractionally integrated ARIMA(p,d,q) model? I've written a program to forecast fractioanlly integrated models that uses the GPH.SRC proc to estimate "d", but would love to integrate a MLE proc into the forecasting proc. --Norm Norman J. Morin nmorin@weber.ucsd.edu ------------------------------------------------------------- Department of Economics * University of California, San Diego 9500 Gilman Drive * La Jolla, CA 92023-0508 ------------------------------------------------------------- ---------- End of message ---------- From: fiorentini@decon.unipd.it (Riccardo Fiorentini) To: "RATS Discussion List" Subject: Re: RATS for Macintosh Date: Thu, 12 Sep 96 11:30:04 +0100 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Organization: dip. scienze economiche X-Mailer: ExpressNet/SMTP v1.1.5 (via Mercury MTS (Bindery) v1.30) MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit pgallagh@ozemail.com.au,internet writes: > Any news yet on the long overdue power mac native version? I recently received an answer from Estima from which it is clear that Mac Rats is not a priority for Estima. However, I think this is not a good policy since other major econometric software like TSP and SHAZAM is already avalaible in PowerMac native version (without any delay!) and works fine. I think this situation may result in customers switching from Rats to other econometric programs. Estima should evaluate this potential loss of customers! Riccardo Fiorentini Department of Economics University of Padova - Italy http://www.decon.unipd.it e-mail fiorentini@decon.unipd.it ---------- End of message ---------- From: jboxstef@magnus.acs.ohio-state.edu (Janet M. Box-Steffensmeier) To: "RATS Discussion List" Subject: Re: Fractional Integration Date: Thu, 12 Sep 1996 08:57:49 -0400 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Mailer: (via Mercury MTS (Bindery) v1.30) David Montgomery has been making great progress on this front. Check out his web page, which contains his programs in progress. His email address is: dmontgom@hawaii.edu Cheers, Janet >Has anyone yet programmed Sowell's maximum likelihood estimation >of a fractionally integrated ARIMA(p,d,q) model? I've written >a program to forecast fractioanlly integrated models that uses >the GPH.SRC proc to estimate "d", but would love to integrate >a MLE proc into the forecasting proc. > >--Norm > > Norman J. Morin > nmorin@weber.ucsd.edu >------------------------------------------------------------- >Department of Economics * University of California, San Diego > 9500 Gilman Drive * La Jolla, CA 92023-0508 >------------------------------------------------------------- > > > > ____________________________________________________________________________ Janet M. Box-Steffensmeier PHONE: (614) 292-9642 or (614) 292-2880 Department of Political Science FAX: (614) 292-1146 Ohio State University E-MAIL: jboxstef@magnus.acs.ohio-state.edu 2140 Derby Hall 154 N. Oval Mall Columbus, OH 43235 ______________________________________________________________________________ ---------- End of message ---------- From: "Christopher F. Baum" To: "RATS Discussion List" Subject: Re: RATS for Macintosh Date: Thu, 12 Sep 1996 09:49:42 -0400 (EDT) Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: MIME-version: 1.0 Content-type: TEXT/PLAIN; CHARSET=US-ASCII Content-transfer-encoding: 7BIT X-Mailer: Mercury MTS (Bindery) v1.30 I agree wholeheartedly. We have mostly Macs in our department, many of them PowerMacs, but most RATS use is via the UNIX or DEC Alpha environments due to instability of MacRATS, and its inability to work with the PowerMac math chip. This is a real Achilles' heel, and encourages many of my colleagues and grad students to look elsewhere for statistical software. Add to the list of PowerMac compatible software SAS 6.10 and Stata, both of which work quite well in the PowerMac environment on large data sets, and whose graphics are at least as impressive as RATS. I'd love to do more of my statistical computing on my desktop, but Estima's inability to get on the stick here is hindering that, and I think hurting them. To dispel the "but no one is developing for the Mac" I would point out Scientific Word / Scientific Workplace (which run ONLY on the PowerMac, not the 68K), for which I am a betatester. TCI has put significant resources into developing these apps; SW is commercial, SWP is still in beta for PowerMac, but they obviously have made the decision that PowerMac is worth their development dollars. And when Mathematica version 3 comes out, guess which platform it will run on firstÉ Kit Baum Boston College baum@bc.edu ---------- End of message ---------- From: Gregory Lypny To: "RATS Discussion List" Subject: Rats for PowerMacs Date: Thu, 12 Sep 96 11:26:38 -0500 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: x-mailer: Claris Emailer 1.1 (via Mercury MTS (Bindery) v1.30) Mime-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" I'm throwing my two cents into the discussion on Rats for PowerMacs. I am very disappointed like the others. Having to run some fairly large GARCH models, among others, and having traded up to the PowerMac at least in part for that purpose, I was forced to buy TSP 4.3, which is native for the PowerMac. TSP will estimate my benchmark GARCH model in about 2 minutes where it takes MacRats about 60 minutes on my old 40 mHz Mac. The trouble is that TSP is not very interactive at all, the example routines are not well "commented", and the manuals do not contain many examples; Rats, in my opinion, scores higher on all of these. I would gladly snap up a PowerMac version of Rats if it were made available before I die. I think Estima is missing out on a big opportunity by not supporting PowerMac users, and by not using the Mac platform to make Rats a truly graphical econometrics package. I would suggest that the people at Estima take a look at some of the featues of HiQ by National Instruments, particularly, the use of graphical icons to represent all objects (scalars, matrices, program code, even graphs). Incorporate some of that stuff into a native PowerMac product, and Estima would blow the others out of the water. Right now, of course, I'd be happy with any PowerMac version. Gregory J. Lypny Associate Professor of Finance Director, Centre for Instructional Technology Mailing address: Contact: ------------------------------------------------------------------- Finance Department 514.848.2926 (voice) Concordia University 514.848.4500 (fax) 1455 De Maisonneuve Boulevard West LYPNY@VAX2.CONCORDIA.CA Montreal, QC H3G 1M8 ---------- End of message ---------- From: "Christopher F. Baum" To: "RATS Discussion List" Subject: Re: Rats for PowerMacs Date: Thu, 12 Sep 1996 17:28:31 -0400 (EDT) Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: MIME-version: 1.0 Content-type: TEXT/PLAIN; CHARSET=US-ASCII Content-transfer-encoding: 7BIT X-Mailer: Mercury MTS (Bindery) v1.30 To add another two cents, Estima should recognize that the _real_ competition is not TSP, SAS etc. but GAUSS and MATLAB. The beauty of RATS relative to most of the traditional stat packages is its extensibility and programming language. One of the real weaknesses is nonlinear optimization, with and without constraints, at which RATS is not at all state-of-the-art. To the extent that more and more people need those latter features, and need to "roll their own" to do the complex estimation tasks, RATS is becoming marginalizedÉGAUSS with its max.likelihood and constrained max.lik. extensions is attracting far more attention from "serious" econometricians. But even applied people like me are wonderingÉwhy is RATS on _any_ platform so slow at nonlinear optimization? I can get far better performance out of something that can be compiled (like MATLAB procedures, or FORTRAN / C code). GAUSS is not a Macintosh app and shows no sign of moving in that directionÉbut to fight a war you have to grab uncontested territory on all fronts. Estima has an opportunity to be a far more powerful stat.package for desktop PowerMacs and Pentiums than the TSP, etc. competition, and if they want to fend off people moving to GAUSS and/or MATLAB, they need to consider that more seriously. The obsolecence of the current Macintosh version is a crying shame. Kit Baum Boston College ---------- End of message ---------- From: Economic Group RBA To: "RATS Discussion List" Subject: RATS for Mac Date: Fri, 13 Sep 1996 08:13:22 +1000 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Mailer: Mercury MTS (Bindery) v1.30 Rob, et al, I had thought that we were "alone in the wilderness", but maybe Estima will get the hint. I know when we upgrade to Mathematica 3 we will be seriously looking at moving off RATS (on both platforms). I know the RBA isn't a large customer, quantity wise, but I would have thought that loosing a central bank as a client would have some effect. Thanks for the support. Mark :-{) Economic Analysis Department Email enquiries: Economic Group PH: +612 9551 8817 Reserve Bank of Australia FAX: +612 9551 8000 ---------- End of message ---------- From: Norman Morin To: "RATS Discussion List" Subject: RATS Wishlist Date: Thu, 12 Sep 1996 15:57:58 -0700 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Mailer: Mercury MTS (Bindery) v1.30 If you were to contribute to a wishlist for future RATS releases what would you ask for? I'd like to see (1) The ability to use a declared vector or matrix as the argument in a NONLIN command. It'd be fairly easy, for example, to then write a generic GARCH(p,q) program. (2) A Power PC native version for the Mac (or, at least, the newest version for the Mac) (3) Faster optimization routines I don't want to seem like I am too critical of RATS -- its combination of canned commands and procedures and its flexibility in writing one's own procedures is without peer. But, like Apple has done, I hope Estima isn't too slow to take advantage of its relative and absolute strengths. The huge increase in processor speeds, too, can allow RATS to be more mouse-menu-icon driven without sacrificing much performance, I should think. Oh, well, back to work. --Norm Norman J. Morin nmorin@weber.ucsd.edu ------------------------------------------------------------- Department of Economics * University of California, San Diego 9500 Gilman Drive * La Jolla, CA 92023-0508 ------------------------------------------------------------- ---------- End of message ---------- From: "Ulrich Leuchtmann" To: "RATS Discussion List" Subject: RECTANGULAR[FRML] Date: Fri, 13 Sep 1996 09:42:38 GMT+0100 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Organization: Fak. WIWI Uni Bielefeld X-Mailer: Pegasus Mail for Windows (v2.23DE) (via Mercury MTS (Bindery) v1.30) According to the RATS User's Manual you can make aggregates of aggregate data types, eg. VECTOR[SYMMETRIC[REAL]]. Within a dynamic input-output-model I tried to work with the data types VECTOR[FRML] and RECTANGULAR[FRML]. Declaration was possible, but all attempts to work with such variables led to an immediate crash of RATS (Vers. 4.*). I also tried the more explicit declarations VECTOR[FRML[REAL]]] and RECTANGULAR[FRML[REAL]]], but that led to the same results. Has somebody else tried these types? If they aren't possible, I'd like to add them to the RATS-Wishlist, which Norman mailed yesterday. Estima should add a note to the User's Manual, if some combinations of aggregate data types aren't possible (p.4-10: "You can make an aggregate of any data typeÉ"). Uli ============================================================= Ulrich Leuchtmann University of Bielefeld German Statistical Society Faculty of Economics Secretariat P.O.Box 10 01 31 Tel.: (+49)(0) 521-106-4870 33501 Bielefeld Fax: (+49)(0) 521-106-2994 Germany email: ULeucht@wiwi.uni-bielefeld.de ============================================================= ---------- End of message ---------- From: Rob Trevor To: "RATS Discussion List" Subject: Re: Wish List & RECTANGULAR[FRML] Date: Fri, 13 Sep 1996 17:55:53 +1000 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Mailer: Mercury MTS (Bindery) v1.30 At 7:42 PM +1000 13/9/96, Ulrich Leuchtmann wrote: > É > Within a dynamic input-output-model I tried to work with the data > types VECTOR[FRML] and RECTANGULAR[FRML]. Declaration was possible, > but all attempts to work with such variables led to an immediate > crash of RATS (Vers. 4.*). > É. > Has somebody else tried these types? If they aren't possible, I'd > like to add them to the RATS-Wishlist, which Norman mailed yesterday. > É As far as I know you can't do it in v4.2. You can pull some tricks with arrays on the RHS of (scalar) FRML statements, but you can't have arrays of FRML's. (See my multivariate GARCH code on the Estima web site for examples.) I too would like to see this in a the next version - as well as user access to the numercial derivatives of the FRML's. (RATS calculates these, you just can't get access to them.) Cheers Rob ----------------------------------------------------------------------- Rob Trevor Associate Professor Centre for Studies in Money, Banking and Finance Macquarie University, NSW, AUSTRALIA 2109 Internet: robt@efs.mq.edu.au VOICE: +61 (2) 9850-8447 FAX: +61 (2) 9850-7281 or +61 (2) 9418-1847 ----------------------------------------------------------------------- ---------- End of message ---------- From: Sune Karlsson To: "RATS Discussion List" Subject: Re: RECTANGULAR[FRML] Date: Fri, 13 Sep 1996 12:48:02 +0100 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Organization: Stockholm School of Economics MIME-version: 1.0 X-Mailer: Mozilla 3.0 (WinNT; I) (via Mercury MTS (Bindery) v1.30) Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7bit Ulrich Leuchtmann wrote: > Has somebody else tried these types? If they aren't possible, I'd > like to add them to the RATS-Wishlist, which Norman mailed yesterday. > Estima should add a note to the User's Manual, if some combinations > of aggregate data types aren't possible (p.4-10: "You can make an aggregate > of any data typeÉ"). Having run in to the same problem (and it is definitely on *my* wish list) I proceeded to read a little further on the page: "É any data type, subject to the following restrictions: É. * Currently, the only type of FRML permitted is FRML[REAL]É" /Sune -- Sune Karlsson | stsk@hhs.se Fax: + 46 8 34 81 61 Stockholm School of Economics | Phone: + 46 8 736 92 39 Box 6501, 113 83 Stockholm, Sweden | http://www.hhs.se/stat/ ---------- End of message ---------- From: "Ulrich Leuchtmann" To: "RATS Discussion List" Subject: Re: RECTANGULAR[FRML] Date: Fri, 13 Sep 1996 13:13:21 GMT+0100 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Organization: Fak. WIWI Uni Bielefeld X-Mailer: Pegasus Mail for Windows (v2.23DE) (via Mercury MTS (Bindery) v1.30) Hi Sune, thanks for your answer: > I proceeded to read a little further on the page: > "É any data type, subject to the following restrictions: > É > * Currently, the only type of FRML permitted is FRML[REAL]É" I don't wanna be insisting, but let me explain, how I understood the User's Manual: The restriction, that the only subtype of FRML is REAL (3rd item on ^^^ the list you quoted (p.4-10)) doesn't rule out the possibility, that a FRML itself can be aggregated (as stated for the aggregate data types in general in the 2nd item of the same list!). Note that, in analogy, the only types of SERIES are SERIES[REAL] and SERIES[COMPLEX], but that a VECTOR[SERIES] is possible. Note also, that in the declaration VECTOR[FRML[REAL]] the 3rd restriction is obeyed explicitely, but that this data type doesn't work. So the impossibility of a VECTOR[FRML] is in contrast with the 2nd item of this list. The suggestion I wanted to make, was, that this inconsistency should be noted in future releases of the Manual. Uli ============================================================= Ulrich Leuchtmann University of Bielefeld German Statistical Society Faculty of Economics Secretariat P.O.Box 10 01 31 Tel.: (+49)(0) 521-106-4870 33501 Bielefeld Fax: (+49)(0) 521-106-2994 Germany email: ULeucht@wiwi.uni-bielefeld.de ============================================================= ---------- End of message ---------- From: Wai Lee To: "RATS Discussion List" Subject: -No Subject- Date: 13 Sep 96 9:22:48 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Mime-Version: 1.0 Content-Type: Text/Plain X-Mailer: Mercury MTS (Bindery) v1.30 The discussion seems to have been expanded to "wish list" rather than just on the demand of RATS for Power Macs. I absolutely agree with Norman Morin on the strength of RATS, and want to add that RATS provides good technical support, but still wish that the followings will be improved soon and why: 1. I may just have bad luck as I have a lot of crash problems with WinRATS (I posted a message here about a month ago asking for help). It crashes with floating point error (from the responses I received, 3 others have had such experience) in MAXIMIZE, crashes with some uses of mouse (9 others have had that), and I just discovered that it will crash if the MAX and MIN options in GRAPH are NA (my 6 hours of work on rolling sample GARCH modeling was gone!). To play save, now I always save the output series immediately after estimations. 2. Occasionally, I find that MAXIMIZE is "extremely" sensitive to startying values, though it works fine generally. Here is my experience. I estimate a simple GARCH(1,1) model with starting values of the (1,1) parameters being 0.1 and 0.8, 0.2 and 0.7, 0.3 and 0.6 ÉÉ. to 0.8 and 0.1 (sort of a grid search). All converge to a point estimate of -0.17 for one parameter, Then I use EZARCH (a less known software that just estimates GARCH, which, É. "when BHHH fails to find a step size that increases the likelihood value, uses a scaled-sown score vector as the direction vector is applied") to verify. It converges to about 0.745 with automatically chosen starting values, and is at least 5 times faster than RATS in this job. Then I play with RATS again with starting value of 0.75. It still failed! I finally got conergence with starting value of 0.741. I have had many other examples like this one from my rolling sample estimations for many assets. I think it is rare to see any software that clearly dominates all others, and same applies here. I use MATLAB or IML of SAS especially for GMM estimations when you have different instruments for each equation, and a few other nonlinear modeling. In terms of speed, for the same nonlinear model I find that RATS is faster than PC-SAS and PC-MATLAB (of course it depends on your own code in SAS and MATLAB), run on Pentium. But, I am seriously considering switching because of its starting value problem, or at least verify its results with some other software. Wai Lee ---------- End of message ---------- From: "Estima" To: "RATS Discussion List" Subject: Many topics Date: Fri, 13 Sep 1996 16:39:21 -0600 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT X-mailer: Pegasus Mail for Windows (v2.31) (via Mercury MTS (Bindery) v1.30) Gee, the list sure has been busy latelyÉ. I figured one big message was easier for everyone to deal with than several short ones, so I'll try to address several issues that have come up this week: 1) VECTOR's of FRML's haven't quite been implemented the way we'd like them to be. In general, they should work well, although I have seen a couple of situations that cause crashes. If you have a program that isn't working, I'd be happy to take a look at it to see if we can fix the problem. 2) Allowing VECTORS as NONLIN parameters has long been on our own wish list as well. We hope to implement them by Version 5.0, if not sooner. 3) Regarding Mr. Lee's problems with MAXIMIZE crashingÉIs it possible that you are running the 16-bit version (WinRATS), rather than WinRATS-32? Please check this carefully. I haven't been able to reproduce any problems with WinRATS-32, but there are a couple of bugs that cause the 16-bit version (and sometimes the DEC Alpha version) to behave the way you've described when encountering divide-by-zero situations. 4) Re. the Mac version--I've passed all the recent comments on to Mr. Doan. As far as I know, Mac updates are still in the pipeline, but I won't hazard a guess as to when they will appear. I mostly wanted to let you know that we are indeed listening, but I can't resist a few comments: Personally, I would have been happy to see the updated Mac versions and PowerMac version ship a long time ago (for many reasons), but I also understand why it hasn't happened yet. Basically, the Mac updates/PowerMac version were back-burnered while we did the Windows versions and DOS/UNIX updates, etc. Macs have always represented a fairly small portion of our user base, and thus don't rate as high on the list of priorities. My understanding is that some of our competitors have had trouble justifying their Mac development costs as well. Unfortunately, the Mac apparently hasn't penetrated the econometrics market very well. Things may have improved recently, with PowerMac proving to be a good number-crunching platform, but I'm not sure about that. And yes, I do know this is a vicious circle--the Mac segment isn't going to grow at all unless we put out good new products, publicize them, etc. We do still sell a fair number of copies of MacRATS, though, so I think there is still justification for continuing to support the Mac platform. Frankly, Apple's recent problems haven't helped the situation. They've been showing signs of life lately, however, which does help. My point? Well, I do hope the MacRATS users out there do stick with us, and I hope that we can deliver some products for you soon. Also, we certainly do appreciate your comments, and would be happy to hear from other past, present, or potential MacRATS users. Thanks, Tom Maycock Estima -- +-----------------------------+-----------------------------------------+ | Estima | | | P.O. Box 1818 | Voice: (847) 864-8772 | | Evanston, IL 60204-1818 | Fax: (847) 864-6221 | | U.S.A | BBS: (847) 864-8816 | | e-mail: estima@estima.com | CompuServe: 73140,2202 | |-----------------------------------------------------------------------| | Web Site: http://www.estima.com | | RATS Internet Mailing List: New members can join by sending e-mail to | | MAISER@EFS.MQ.EDU.AU with the message: SUBSCRIBE RATS-L | +-----------------------------------------------------------------------+ ---------- End of message ---------- From: Wai Lee To: "RATS Discussion List" Subject: Unix RATS and Web Date: 18 Sep 96 13:17:46 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: Mime-Version: 1.0 Content-Type: Text/Plain X-Mailer: Mercury MTS (Bindery) v1.30 Has anyone used RATS for Unix with Web? Can I type in input, click on a button on a web page and run RATS at the background, and then deliver the results back to web? How should the CGI be done in this case? Thanks for any advice. Wai Lee JPMIM ---------- End of message ---------- From: "Estima" To: "RATS Discussion List" Subject: Re: Unix RATS and Web Date: Wed, 18 Sep 1996 17:21:08 -0600 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT X-mailer: Pegasus Mail for Windows (v2.31) (via Mercury MTS (Bindery) v1.30) > Has anyone used RATS for Unix with Web? Can I type in input, click on a button > on a web page and run RATS at the background, and then deliver the results back > to web? How should the CGI be done in this case? > > Thanks for any advice. > > Wai Lee > JPMIM I do actually know of one user who has developed a Web-based interface for UNIX RATS. It allows you to type in and run new programs, or select programs from a drop-box, edit them, and run them. Output is displayed to a new page, and any graphs are converted to GIF format and displayed as thumbnails on the output page. I've tried it myself, and it works very well. I will contact the user to see if he wants to share his work. I'll report back to the list when I hear anything. Thanks, Tom Maycock Estima -- +-----------------------------+-----------------------------------------+ | Estima | | | P.O. Box 1818 | Voice: (847) 864-8772 | | Evanston, IL 60204-1818 | Fax: (847) 864-6221 | | U.S.A | BBS: (847) 864-8816 | | e-mail: estima@estima.com | CompuServe: 73140,2202 | |-----------------------------------------------------------------------| | Web Site: http://www.estima.com | | RATS Internet Mailing List: New members can join by sending e-mail to | | MAISER@EFS.MQ.EDU.AU with the message: SUBSCRIBE RATS-L | +-----------------------------------------------------------------------+ ---------- End of message ---------- From: "Estima" To: "RATS Discussion List" Subject: Re: Unix RATS and Web Interface Date: Thu, 19 Sep 1996 17:08:00 -0600 Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT X-mailer: Pegasus Mail for Windows (v2.31) (via Mercury MTS (Bindery) v1.30) Dear folks: Further to my earlier message about the Web-based interface, I've talked (well, e-mailed) with one of the guys working on this. They are going to do some further work in preparation for the start of classes in late September. At that point, he'll get back to me with their ideas on distributing their stuff. Thanks, Tom Maycock Estima -- +-----------------------------+-----------------------------------------+ | Estima | | | P.O. Box 1818 | Voice: (847) 864-8772 | | Evanston, IL 60204-1818 | Fax: (847) 864-6221 | | U.S.A | BBS: (847) 864-8816 | | e-mail: estima@estima.com | CompuServe: 73140,2202 | |-----------------------------------------------------------------------| | Web Site: http://www.estima.com | | RATS Internet Mailing List: New members can join by sending e-mail to | | MAISER@EFS.MQ.EDU.AU with the message: SUBSCRIBE RATS-L | +-----------------------------------------------------------------------+ ---------- End of message ---------- From: "L.Saldanha" To: "RATS Discussion List" Subject: Switching regressions Date: Tue, 24 Sep 1996 10:19:58 +0100 (BST) Errors-to: Reply-to: "RATS Discussion List" Sender: Maiser@efs1.efs.mq.edu.au X-listname: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Mailer: Mercury MTS (Bindery) v1.30 Hi, I am currently doing some work with switching regressions and was wondering if somebody could tell me if there is any econometric package which provides a general maximum likelihood procedure if switching points are unknown Thank You. L. Saldanha Research Student ---------- End of message ----------