| DEMO_SVMPOUT | |
Fitting a posteriory probability to SVM output.
A posteriory probability p(y==1|f(x)) of the first class
given SVM output f(x) is assumed to be sigmoid function.
Parameters A(1) and A(2) of the sigmoid function
p(y==1|f(x)) = 1/(1+exp(A(1)*f(x)+A(2))
are estimated using Maximum-Likelihood [Platt99a].
The Gaussian mixture model (GMM) is fitted to the SVM output
and the a posteriory probability are computed for
comparison to the ML estimate.
The ML estimation of the sigmoid function is imlemented
in 'mlsigmoid' (see 'help mlsigmoid' for more info).
About: Statistical Pattern Recognition Toolbox
(C) 1999-2003, Written by Vojtech Franc and Vaclav Hlavac
Czech Technical University Prague
Faculty of Electrical Engineering
Center for Machine Perception
Modifications:
03-jun-2004, VF
6-May-2003, VF