Code for multi-task boosting using regression trees as weak learners. 
The main functionality is implemented in Booster.h and RegressionTree.h
(in the gBoost directory). RegressionTreeLearn.cpp and TaskBoostLearn.cpp
provide mex functions. runBooster.m shows how these can be used in MATLAB.

To compile the code in gBoost, you will need the following libraries:
- Eigen (http://eigen.tuxfamily.org)
- Boost (http://www.boost.org/).
This code was tested with Boost version 1.54.

Then, go to the gBoost directory, modify the paths to these libraries in 
the Makefile, and type make.
 
This script was used to generate the results in:

"Modeling context-specific gene regulation with multi-task boosting", 
Sofia Kyriazopoulou-Panagiotopoulou, Marco Cusumano-Towner, 
Serafim Batzoglou and Anshul Kundaje, 
NIPS Workshop in Machine Learning in Computational Biology, 2013.