PSVM |
Plots decision boundary of binary SVM classifier.
Synopsis:
h = psvm(...)
psvm(model)
psvm(model,options)
Description:
This function samples the Support Vector Machiones (SVM) decision
function f(x) in 2D feature space and interpolates isoline
width f(x)=0. The isolines f(x)=+1 and f(x)=-1 are plotted as well.
Input:
model [struct] Model of binary SVM classifier:
.Alpha [1 x nsv] Weights of training data.
.b [real] Bias of decision function.
.sv.X [dim x nsv] Support vectors.
.options.ker [string] Kernel function identifier.
See 'help kernel' for more info.
.options.arg [1 x nargs] Kernel argument(s).
options [struct] Controls apperance:
.background [1x1] If 1 then backgroud is colored according to
the value of decision function (default 0).
.sv [1x1] If 1 then the support vectors are marked (default 1).
.sv_size [1x1] Marker size of the support vectors.
.margin [1x1] If 1 then margin is displayed (default 1).
.gridx [1x1] Sampling in x-axis (default 25).
.gridy [1x1] Sampling in y-axis (default 25).
.color [int] Color of decision boundary (default 'k').
Output:
h [struct] Handles of used graphical objects.
Example:
data = load('riply_trn');
model = smo( data, struct('ker','rbf','arg',1,'C',10) );
figure; ppatterns(data);
psvm( model, struct('background',1) );
See also
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:
25-may-2004, VF
10-may-2004, VF
5-oct-2003, VF, returns handles
14-Jan-2003, VF
21-oct-2001, V.Franc
16-april-2001, V. Franc, created