RoBO uses the Gaussian processes library george and the random forests library pyrfr. In order to use these libraries make sure that libeigen and swig are installed:
sudo apt-get install libeigen3-dev swig
Download RoBO and then change into the new directory:
git clone https://github.com/automl/RoBO
cd RoBO/
Install the required dependencies.
for req in $(cat requirements.txt); do pip install $req; done
Finally install RoBO by:
python setup.py install
You can find the documentation for RoBO here http://automl.github.io/RoBO/
To cite RoBO please reference our BayesOpt paper:
@INPROCEEDINGS{klein-bayesopt17,
author = {A. Klein and S. Falkner and N. Mansur and F. Hutter},
title = {RoBO: A Flexible and Robust Bayesian Optimization Framework in Python},
booktitle = {NIPS 2017 Bayesian Optimization Workshop},
year = {2017},
month = dec,
}