From: Theodore Papadopoulo (Theodore.Papadopoulo_at_[hidden])
Date: 2006-10-06 05:39:22
On Fri, 2006-10-06 at 10:04 +0100, John Maddock wrote:
> > how about sigmoid, double sigmoid and their derivatives? Used for
> > neural networks.
> Those are new ones to me.
Well, this not so strange. As far as I know, the term sigmoid does not
describe a function but merely a family of functions of a similar shape
(the actual equation is not so important). Eg 1+tanh(x). IMHO unless
there is a more formal or accurate concept of what is a sigmoid, it
should not be added to the library.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk