Author

Sean Holden
Computer Laboratory
University of Cambridge

Quick guide

• You'll need to go to the c folder and do an R CMD SHLIB outer.c. The C code is only needed to speed up computing of the kernel matrices. At present this only implements RBFs. kernels-c.R encapsulates this.

• If you do source("demo.R") then demo4P() you should get a simple example of a problem with 4 classes learned using probabilistic outputs. easyData.R is a bunch of stuff for making trivial data sets for testing.

• cv.R is for stratified cross-validation and is needed because it forms part of the method for turning outputs into probabilities.

• multiShared.R is mostly stuff for transforming multi-class problems into sets of binary problems.

• You then have a whole bunch of stuff named lpSVM*-c.R If the name includes Multi it's for multiple classes, otherwise it's binary. P denotes probabilistic outputs and 2K indicates it's the 2-kernel version I developed to combine LOPIT and GO.

• In the non-2K code data is supplied as a list(X, y) with one example per row in X. y is +/-1 for binary problems and positive integers (preferably 1:n) for n classes.

• In the 2K versions there is in addition X2, which is the features for the auxilliary data-only the primary data should be in X. Also in the routines for classifying new stuff you have x for the primary data and x2 for the auxilliary data. Thus c(x, x2) is the actual new vector you're classifying.

LICENSE

This code is licensed under GNU General Public License, version 2 (GPL-2.0). A copy of the license if distributed with the code.

    Linear Programming SVM Transfer Learning
    https://github.com/ComputationalProteomicsUnit/lpsvm-tl-code
    Copyright (C) 2015  Sean Holden

    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation; either version 2 of the License, or
    (at your option) any later version.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License along
    with this program; if not, write to the Free Software Foundation, Inc.,
    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.