A Semantic Similarity Model for Mapping Between Evolving Geospatial Data Cubes

In a decision-making context, multidimensional geospatial databases are very important. They often represent data coming from heterogeneous and evolving sources. Evolution of multidimensional structures makes difficult, even impossible answering to temporal queries, because of the lack of relationships between different versions of spatial cubes created at different time. This paper proposes a semantic similarity model redefined from a model applied in the ontological field to establish semantic relations between data cubes. The proposed model integrates several types of similarity components adapted to different hierarchical levels of dimensions in multidimensional databases and also integrates similarity between features of concepts. The proposed model has been applied to a set of specifications from different inventory in Montmorency Forest in Canada. Results show that the proposed model improves precision and recall compared to the original model. Finally, further investigation is suggested in order to integrate the proposed model to SOLAP tools as future works.