Evolving Explanatory Novel Patterns for Semantically-Based Text Mining - Natural Language Processing and Text Mining - page 163

IF goal(perform(19311)) and goal(analyze(20811))
THEN establish(111)

Where the numerical values represent internal identifiers for the arguments and
their semantic vectors, and its resulting criteria vector is
[ 0 . 92 , 0 . 09 , 0 . 50 , 0 . 005 , 0 . 7 , 0 . 00 , 0 . 30 , 0 . 25] (the vector's elements represent the val-
ues for the criteria relevance, structure, coherence, cohesion, interestingness, plau-
sibility, coverage, and simplicity) and obtained an average expert's assessment of
3 . 74. In natural-language text, this can roughly be interpreted as (each item of
the following NL description represents a predicate-level information of hypothesis
above):

• IF the work aims at performing the genetic grouping of seed populations and

investigating a tendency to the separation of northern populations into different
classes, AND

• The goal is to analyse the vertical integration for producing and selling Pinus

Timber in the Andes-Patagonia region.

• THEN as a consequence , the best agricultural use for land lots of organic

agriculture must be established to promote a conservationist culture in priority
or critical agricultural areas.

The hypothesis appears to be more relevant and coherent than others (relevance
= 92%). However, this is not complete in terms of cause-effect. For instance, the
methods are missing. It is also important to highlight that the high value for the
coherence of the pattern (50%) is consistent with the contents of the predicates of the
hypothesis. The three key paragraphs containing rhetorical knowledge indeed relate
to the same topic: testing and producing specific Pinus trees. Even more important
is the fact that despite having zero plausibility (novelty), the pattern is still regarded
as interesting by the model (70%) and above the average by the experts. As for the
target concepts ( degradant and erosive ) and the way the discovered hypothesis
attempts to explain the link between them, it can be seen that the contents of this
patterns try to relate these terms with ”agricultural areas,” ”seed populations,” etc.,
so the discovery makes a lot of sense.
Another of the discovered patterns is given by hypothesis 88 of run 3, which is
represented as follows:

IF goal(present(11511)) AND
method(use(25511))
THEN effect(1931,1932)

and has a criteria vector [0 . 29 , 0 . 18 , 0 . 41 , 0 . 030 , 0 . 28 , 0 . 99 , 0 . 30 , 0 . 50] and obtained
an average expert's assessment of 3 . 20. In natural-language text, this can roughly
be interpreted as:

• IF the goal is to present a two-dimensional scheme for forest restoration in

which two regression models with Pinus and without Pinus are identified by
inspiring in the natural restoring dynamics, AND

• The method is based on the use of micro-environments for capturing the kind

of farm mice called Apodemus Sylvaticus i, and on the use of capture traps at a
rate of 1464 traps per night.