leastSquaresFit

[Fernando Pérez]

Alex Tournier wrote:

> Fernando Pérez <fperez528 at yahoo.com> wrote in message
>> Many numeric functions do not
>> implicitly convert lists to arrays (though some can operate on lists as
>> well as arrays). Try data = Numeric.array(data) and see what happens.
>> 
>> f
> 
> I tried as you said, putting it through array(), but still get the same
> message! Could there be a problem with the FirstDerivatives.py module??
> Alex

Well, it seems like you need to upgrade numeric or you need to upgrade 
python. On my machine (python 2.2, Numeric 20.3) it works just fine:


IPython profile: tutorial

*** Pasting of code with ">>>" or "..." has been enabled.

In [1]: >>> from Scientific.Functions.LeastSquares import leastSquaresFit

In [2]: >>> import Numeric

In [3]: >>> # The mathematical model.

In [4]: ... def exponential(parameters, x):
   ...: ...     a = parameters[0]
   ...: ...     b = parameters[1]
   ...: ...     return a*Numeric.exp(-b/x)
   ...:

In [5]: >>> # The data to be fitted to.

In [6]: ... data = [(100, 4.999e-8),
   ...: ...     (200, 5.307e+2),
   ...: ...     (300, 1.289e+6),
   ...: ...     (400, 6.559e+7)]

In [7]:

In [8]: >>> fit = leastSquaresFit(exponential, (1e13, 4700), data)

In [9]: fit
Out[9]: ([8641551709749.7666, 4715.4677901570467], 1080.2526437958597)

cheers,

f.