Two-dimensional binary classification example.
demogp_class
- demo script for using the OGP package for
classification.
The script generates data from a mixture of three Gaussians with two of
them belonging to 'class1'
and the third to 'class2'
.
After the generation of the training data, iterates the computation of the posterior (with the sparse online learning algorithm) and the adjustment of the kernel hyperparameters in the OGP (net).
Similar to regression (demogp_reg
), the script computes and displays
average test errors for different BV set sizes.
The kernel family and the number of iteratins have to be set manually in the
script - see demogp_class_gui
for a more interactive classification demo.
ogp
, ogptrain
, c_class_bin
, demogp_class_gui
, demogp_reg
Copyright (c) Lehel Csató (2001-2004)