This paper presents a new model for flexible noun phrase detection, which is able to recognize
noun phrases similar enough to the ones given by the inferred noun phrase grammar. To allow
this flexibility, we use a very accurate set of probabilities for the transitions between the
part-of-speech tag sequence which defines a noun phrase. These accurate probabilities are
obtained by means of an evolutionary algorithm, which works with both, positive and negative
examples of the language, thus improving the system coverage, while maintaining its precision.
We have tested the system on different corpora and compare the results with other systems,
what has revealed a clear improvement of the performance.