different results from nls in 2.10.1 and 2.11.1
Interesting! I get nice convergence in both 32 and 64 bit systems on 2.13.0. I agree the older versions are a bit of a distraction. The inconsistent behaviour on current R is a concern. Maybe Philip, Uwe, and I (and others who might be interested) should take this off line and see what is going on. I'll offer to act as tabulator of results and have set up a temporary wiki on http://macnash.telfer.uottawa.ca/PBettProb to allow tries to be saved for comparison. Needs a login to "write". Please use username=Ruser password=statistics Hopefully there won't be wikispam, or I'll have to close it off. JN
On 06/20/2011 06:00 AM, r-help-request at r-project.org wrote:
Date: Mon, 20 Jun 2011 11:25:54 +0200 From: Uwe Ligges <ligges at statistik.tu-dortmund.de> To: p.e.bett at dunelm.org.uk Cc: r-help at r-project.org, Philip Bett <pbett at astro.uni-bonn.de> Subject: Re: [R] different results from nls in 2.10.1 and 2.11.1 Message-ID: <4DFF1222.7040202 at statistik.tu-dortmund.de> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Since one is a 32-bit and the other one a 64-bit, and therefore the compiler is also different, you can get different numerical results easily, even with identical versions of R (and most of us do not have outdated R installations around). I just tried your example on 3 different systems with R-2.13.0 and all told me "singular convergence"... Uwe Ligges