Accept certain known extra words in a license without losing score points

When we have a license matched exactly to a rule but there are some extra words, we do not have a 100% match score by design.
Yet there are cases where these extra words are NOT something that should degrade the score. For instance:

`The library is licensed under the GPL` as a rule
vs.
`The FOOBAR library is licensed under the GPL` as a scanned text.

Here the extra `FOOBAR` should not impact the 100% score of the rule detection.

One approach could be to maintain a set of such known words and use these in some post-processing of the license match results to keep a 100% match.