www.tlab.it
Occurrences and
Cooccurrences
The Occurrences, in fact,
are quantities which result from the computation of how many times
(frequences) a single lexical unit
(LU)occurs within a corpus or within the
context units (CU) in which it is
subdivided.
Their distribution can be represented in contingency tables as
follows:
Cooccurrences, then, are
quantities which result from a computation of how many times two or
more lexical units are present together in the same elementary contexts (EC).
Their distribution can be represented in tables such as
the following:
With a simple transformation, the "A" type table
(rectangular) can be transformed into "B" type (squared and
symmetrical) in which for each pair of lexical units the quantity
of their cooccurrences is indicated, that is the total number of
the elementary contexts in which they are present together.
In TLAB text analysis is mostly carried out by
the study of relationships among occurrences and cooccurrences,
either through specific association
indexes, or through the use of multidimensional statistical
techniques like cluster analysis and
correspondence analysis
