Providing Semantic Interpretation to Olap Queries: Application to Hospital Management Data Warehouses

Purpose: The aim of this paper is to present a new methodology to enrich the analysis over data cubes in an OLAP system, allowing the system to add semantic interpretation to the query results adapted to the final user. Methodology: Defining a semantic translation of a given value regarding a concrete range of values can be done by means of fuzzy logic. In the case of querying data cubes, the possible combinations of dimensions, levels and aggregation functions makes this range different for each query, so a static interpretation is not valid. We propose a way to supply a semantic interpretation that adapts itself to the query considering the granularity and the semantic of the used function. Findings: We propose a mechanism that allows the representation of different semantic interpretations in data-cubes, in addition to their adaptation when querying. With it users are automatically provided with personalized and elaborated results that are adapted and really meaningful for their individual requirements. In addition, the mechanism allows an individual (the meaning of a value is independent of the other values) or comparative interpretation (the semantic compare each value with the others in the query). Practical Implications: The accessibility improvements achieved allow the creation of simpler interfaces and improve the query over data-cubes from mobile devices allowing the user to interpret the results without needing to review all the values for the queries. Originality/Value: As far as the authors know, there are proposals to interpret values in a natural way for user, but none of them are able to adapt it to the structure of the query and the functions used, so the user is limited to the considered cases. In our approach, the system will adapt the semantic after each query, providing the user with an easy and interpretable result. Even more, the system can adapt the results to different users in a very simple way.