Multidimensional concept extraction and validation from OLAP requirements in NL

This paper presents a natural language-based approach for requirements specification of OLAP systems. In this approach, requirements are specified through a template representing the concepts of the decision making process. In addition, it provides for the acquisition of analytical queries written according to a linguistic pattern. This latter facilitates the automatic analysis of the queries without being too restrictive to the writing styles. In addition, it allows the coherence verification and validation of the requirements with respect to a given source. Furthermore, it allows the automatic generation of conceptual schemas for the target system, that is mapped to the data source used in the loading step.