mauro.prencipe at unito.it
Wed Oct 28 08:30:44 CDT 2009
I'm currently doing non linear regression calculations by using the lbfgs
package. I've 2 questions:
1) Is there the possibility to get estimations of the errors associated to
each refined parameter (I've experience of other environments like xlispstat
which provide such estimations).
2) If each of the variables used as argument to the lbfgs call is a bfloat
object, so are the functions employed in the fit? I mean:
I've a call like
estimates: lbfgs(FOM, '[K0,V0,kp,e0], ini, eps, [1, 0]);
where the variables occurring in FOM, K0, V0, etc... but eps (I get an error
if I define eps as a bfloat) are defined as bfloat(),fpprec:24$; is the
regression really made with the same precision? I ask this since, actually,
I get troubles when eps (the precision is set to numbers lower than 1.e-16).
More information about the Maxima