Java program to analyse text (in the first instance from bookmarks)

Using code from the books:
collective intelligence in action - ISBN: 1933988312

Adapted to only inject phrases and synonyms (from config files) once.
The original code would inject the phrase, all of its components and all of the synonyms.
So for example 'Collective Intelligence in Action' would inject the phrase, the words Collective, Intelligence, in and Action and 
the synonym ciia.
I also want to avoid stemming/stop word removal too early othrwise the phrase, 'Collective Intelligence in Action' could
match with 'Collects Intelligence for Action'.
That should help in maximising the signal to noise ratio for clustering/classification. 

Note: built with netbeans.