The document titles give an important information about documents. This is why they are frequently used to obtain document
keywords. We use them to determine document intentions. To obtain some textual details, we use special information extraction
techniques for the construction of extra-topical representations of the documents. This representation reflects a document
more completely. A possible use for the representation in the information retrieval is described. This improves the retrieval
results.
This work was done while Manuel Montes y Gómez was supported by CONACYT, Mexico, through scholarship to pursue his Master
Sc. Degree. The work was also partially supported by REDII-CONACYT and DEPI-IPN, Mexico.