Supervised reconstruction of biological networks with local models
Kevin Bleakley, Gérard Biau and Jean-Philippe Vert
Bioinformatics, 23(13):i57-i65, 2007
Contains all data used in the paper, in MATLAB format.
Code
The following MATLAB functions were used in the paper:
- graphinference.m : the main function to perform graph inference and compare different algorithms (check the header of the file for details)
- simpleSVMcustem, a slightly customized version of SimpleSVM. You need to tar xvzf this file and put the content of the foler in the MATLAB path for graphinference.m to work. The only difference with the original simpleSVM is that we added the possibility to use a precomputed Gram matrix with the 'custom' kernel type.
Although we use the simpleSVM implementation of a support vector machine, it should be straightforward to adapt the code to any SVM implementation. You also need the rocanalysis.m and metriclearning.m functions on the path.
Vert Jean-Philippe
Last modified: Mon Nov 26 16:14:21 CET 2007