From cottrell@wfu.edu Wed Dec 22 07:50:53 2010
From: Allin Cottrell <cottrell@wfu.edu>
To: gretl-devel@gretlml.univpm.it
Subject: Re: [Gretl-devel] bad nls error message
Date: Wed, 22 Dec 2010 07:50:52 -0500
Message-ID: <Pine.A41.4.58.1012220745070.549302@f1n11.sp2net.wfu.edu>
In-Reply-To: 26399_1292954940_oBLI906H027184_4D10ED33.9000105@gmx.net
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="===============3434628077633848485=="

--===============3434628077633848485==
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit

On Tue, 21 Dec 2010, Sven Schreiber wrote:

> I just tried to do NLS w/o initializing the parameters -- the error
> message was not informative ("wrong data type"). If I use "catch", I get
> an even stranger message: "no regression function is given".

Sample script, please.

I've tried modifying/breaking the file ects_nls.inp, supplied with
gretl, in two different ways:

* Remove the definition of the parameter "alpha" altogether --
Error message: The symbol 'alpha' is undefined

* Give a bare declaration of alpha, no initialization --
Error message: Data error

Allin

--===============3434628077633848485==--


From svetosch@gmx.net Wed Dec 22 09:59:17 2010
From: Sven Schreiber <svetosch@gmx.net>
To: gretl-devel@gretlml.univpm.it
Subject: Re: [Gretl-devel] bad nls error message
Date: Wed, 22 Dec 2010 15:59:10 +0100
Message-ID: <4D12123E.7080409@gmx.net>
In-Reply-To: Pine.A41.4.58.1012220745070.549302@f1n11.sp2net.wfu.edu
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="===============8760605014280329194=="

--===============8760605014280329194==
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit

Am 22.12.2010 13:50, schrieb Allin Cottrell:
> On Tue, 21 Dec 2010, Sven Schreiber wrote:
> 
>> I just tried to do NLS w/o initializing the parameters -- the error
>> message was not informative ("wrong data type"). If I use "catch", I get
>> an even stranger message: "no regression function is given".
> 
> Sample script, please.
> 
> I've tried modifying/breaking the file ects_nls.inp, supplied with
> gretl, in two different ways:
> 
> * Remove the definition of the parameter "alpha" altogether --
> Error message: The symbol 'alpha' is undefined
> 
> * Give a bare declaration of alpha, no initialization --
> Error message: Data error
> 

<script>
open denmark
#c_1 = 0
#c_2 = 0
#c_3 = 1
nls LRM = c_1 + c_2 * (IBO-c_3*LRY)^2
    deriv c_1 = 1
    deriv c_2 = (IBO-c_3*LRY)^2
    deriv c_3 = -2*c_2*(IBO-c_3*LRY)
end nls
</script>

However, maybe it's just a translation issue, I was re-translating back
from German... but "wrong data type" is definitely the right re-translation.

-sven


--===============8760605014280329194==--


From cottrell@wfu.edu Wed Dec 22 10:48:16 2010
From: Allin Cottrell <cottrell@wfu.edu>
To: gretl-devel@gretlml.univpm.it
Subject: Re: [Gretl-devel] bad nls error message
Date: Wed, 22 Dec 2010 10:48:12 -0500
Message-ID: <Pine.A41.4.58.1012221046550.532912@f1n11.sp2net.wfu.edu>
In-Reply-To: 4D12123E.7080409@gmx.net
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="===============0699801279699954814=="

--===============0699801279699954814==
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit

On Wed, 22 Dec 2010, Sven Schreiber wrote:

> Am 22.12.2010 13:50, schrieb Allin Cottrell:
> > On Tue, 21 Dec 2010, Sven Schreiber wrote:
> >
> >> I just tried to do NLS w/o initializing the parameters -- the error
> >> message was not informative ("wrong data type"). If I use "catch", I get
> >> an even stranger message: "no regression function is given".
> >
> > Sample script, please...
>
> <script>
> open denmark
> #c_1 = 0
> #c_2 = 0
> #c_3 = 1
> nls LRM = c_1 + c_2 * (IBO-c_3*LRY)^2
>     deriv c_1 = 1
>     deriv c_2 = (IBO-c_3*LRY)^2
>     deriv c_3 = -2*c_2*(IBO-c_3*LRY)
> end nls
> </script>

Thanks. There's now an improved error message for this sort of
case in CVS.

Allin

--===============0699801279699954814==--