Most of the information that is used as input for the development of information systems is originally produced in non-structured forms, such as verbal communications or free-style text documents. Some examples are requirements specifications or documents associated to translation of contents. In cases like these, there is a need to support the structuring and extraction of the underlying semantic relations embedded in the text, in order to fully understand and post-process it. There are models for the analysis of textual information, such as information retrieval solutions or topic maps. However, these models do not offer an integral, flexible and reusable approach that can assist in giving structure and semantics to the information within a standard framework. Discourse analysis techniques, in contrast, identify semantic relations (such as authors' intentionality or possible dependencies between text elements) in the textual information according to well-established linguistic patterns. Based on these techniques, this paper presents a modelling language based on the ISO/IEC 24744 metamodel that is capable of representing pieces of textual information in a highly structured manner, describing the semantic relations in the associated discourse. In addition, the paper shows an application of the proposed language to the domain of requirements engineering, illustrating the benefits of the application of the suggested approach as well as its possibilities in other textual domains.

2014 IEEE Eighth International Conference on Research Challenges in Information Science, May 28-30 2014, Marrakesh (Morocco).

Peer Reviewed