Created by: anjos

As it is today, this script requires 2 score sets for running (dev and test).

It is possible to compute the performance using a single set.

Here are some suggestions for improvements:

1. Make the dev score set an argument instead of an option. It is strange conceptually to have obligatory "options"
2. Make the test score set an optional argument (this implies in allowing the analysis to run if only dev is provided