Collocation

Collocation setting

• Node

  • Insert search term here, for example 'take' (see Query)

  • Span = number of words in left and right context

  • min-cooccurence = minimal observed frequency within the span

  • max collocates (maximum 

• Result presentation

  • Table, Charts and Network Graph

  • charts and network graphs = all, left dominant, right dominant collocates

Filter search

• By Token/POS/Lemma filter

• • Filter by token/POS/lemma (only collocates that you want to see), to see only adjective, noun etc. You can use minus sign to hide output. for example, when you enter -new and JJ in token and POS fillter, the list will show only adjectives and hide 'new'

• By pattern

  • Filter by Token/POS/Lemma filter is agnostic of position. To specify exactly in which position (Left or right, L1, L2, R1, R2 etc) 

  • Format: label colon formula

    • Agent of passive : <> oleh # 

      • Node: given

      • label = Agent of passive di-

      • Formula = <> by #

        • this means find a collocate (#) preceded by 'by' and the node <#> 

Pattern Search filter

Symbols:

• <> : the node word

• # : the collocate

• * : optional token (any word, 0 or 1)

• + : required token (exactly 1 word)

• token : specific token (obligatory)

• (token) : optional specific token

Constraints:

• [lemma] : token must be from specified lemma

• _TAG : token must have specified POS tag

• (_TAG) : optional POS tag constraint

• ([lemma]) : optional lemma constraint