Corpus preparation

Standard query

• direct token: write the token as it is, as in 'before', 'always', 'developing'

• wildcard * (any string/word)

  • * : find any word (make sure not attach to any string)

  • *ing = find all words ending in -ing

  • un* = find all words starting in un-

  • *zz* = find all words containing 'zz' in it

• POS tag: begin your tag with an underscore

  • • _JJ : find all adjectives

    • _VB : find all verbs

    • ...

• Lemma: sorround the lemma with []

  • [see]: find all words whose lemma is see (see, saw, seeing, sees)

• Multiword search

  • always happy : always + happy

  • always _JJ: always +  happy/sad/angry/beautiful 

  • [develop] experimental _NN: develop/develops/developed/developing experimental drugs/medicine...

Natural Language query: AI  and rule based

• • Instead of writing a search formula like '_JJ _NN' (JJ = adjective, NN=Noun), you can write 'find me an adjective followed by noun'. There is a rule based option too which usually works faster than AI

Restricted search

• Limit your search on certain segment(s) of the corpus

  • e.g. if you choose economy, then only segment(s) from economy domain will be analysed

  • Note that this feature is only active if your corpus has been marked up