Parse tree of the discussion topic. This parse tree has been generated using the Stanford parser:  

Contexts in source publication

Context 1

... identifies the category of each constituent (for instance: determiner, noun, or verb). This parsing process results in a parse tree. Since nouns and noun phrases can be used as key concepts in a discussion topic, we select from the parse tree of the parsed discussion text only constituents which are tagged as nouns (NN) or noun phrases (NP) (cf. Fig. 1). Since the present implementation of our approach is not able to determine which concept is more important than another one. Thus, the system proposed here uses all extracted key concepts that are marked as NN or NP in the resulted parse ...

Context 2

... the context of topic 1, Table 5 shows that the mean of understandability for human-generated questions (2.28) is a little higher than of system-generated questions 19). However, their difference is statistically not significant. ...