Resolution of Demonstrative Anaphoric References in Portuguese Written Texts

This paper presents a proposal for resolving demonstrative anaphoric references on the basis of written Portuguese text corpus. Our study examined the possibility for applying a selection of merely heuristic and syntactical rules into a computational model aimed at the resolutions anaphoric references. Our proposal has been experimented and evaluated: its resolution capacity revealed more than 90% precision with 80% and better recall.