Read Me

Details on the methodology and implmentation can be found on:
R. B. Uriarte, S. Tsaftaris and F. Tiezzi, Service Clustering for Autonomic Clouds Using Random Forest. In Proc. of the 15th IEEE/ACM CCGrid [In Press], 2015.

The RF+PAM algorithm was tested only in Linux and was developed using Python. There's no need to install it, it runs directly from the API or from the RFPAM.py file:

Run the algorithm in the command line and an interactive menu will ask for the datasets:
    python RFPAM.py


Dependencies:
    np - http://www.numpy.org/
    r2py - pip install rpy2  - http://rpy.sourceforge.net/  - In future versions we will remove this dependencies with natives PAM implementations
    R - http://www.r-project.org/

The algorithm can also be used from its API. See the examples in the "Examples" folder. We provide some example datasets in the "Data" folder. They can be used for tests with the algorithm. 

IMPORTANT - The datasets should not have ID of the observation and the features should be separated by ','