The purpose of this file is to outline the procedure of extracting feature set 4.
First those sentences containing Boosters were detected, extracted and exported (2.4.1 Booster Detection)
This subset served as an input for the Stanford Grammatical Parser. The Parser was used to create grammatical dependency tuplets (2.4.2 Booster Grammatical Dependency Analysis)
Using the dependency tuplets, those words being related to a Booster could be detected. Those words were tagged as “_HIGH" (2.4.3 Booster Tagging)
Steps 1-3 were performed for Attenuators.
In order to tag the remaining words, a subset of non-tagged words was created and used as an input for POS Tagging. Finally the text elements were remerged and the final feature set was created.