The point of it...

The idea is to find out how strongly each collocate relates to the search-word near which it was found.

 

How to compute it

In the Concord menu, choose Compute | Relationships:

 

Now choose a word list

The collocation display initially shows the collocates without any relation information (eg. 0.00); this is because to compute the relationship we need to know

a) how often each collocate appears in the corpus we're using,

b) how often the search-word appears in the corpus, and

c) how often they come together within the horizons selected.

The problem is that although b) and c) are known at the time the concordance is computed, a) is not known without doing a concordance or word list for each collocate....

So you are asked to choose an appropriate word list created by WordSmith, which will know the frequencies for each word. It's up to you to choose a word list which actually relates to the concordance you've done!

 

Full lemma processing, case sensitive

These should be checked if your word list has any lemmatised entries, or it is a case-sensitive word list.

 

Type of relationship

Choose in the main Controller Concord settings which type of relation you wish to compute. The default is Specific Mutual Information.

 

Recompute tokens

Choose in the main Controller Index settings whether you want all tokens to be recomputed afresh.

 

 

See also: Collocation, Collocate display, Mutual Information

 

 

Page url: http://www.lexically.net/wordsmith/step_by_step_Chinese/?collocationrelationship.htm