Automatic textual aggregation approach of scientific articles in OLAP context

摘要

In the last decade, Online Analytical Processing (OLAP) has taken an increasingly important role in Business Intelligence. Approaches, solutions and tools have been provided for both databases and data warehouses, which focus mainly on numerical data. These solutions are not suitable for textual data. Because of the fast growing of this type of data, there is a need for new approaches that take into account the textual content of data. In the context of Text OLAP (OLAP on text or documents), the measure can be textual and need a suitable aggregation function for OLAP operations such as roll-up. We present in this paper a new aggregation function for textual data. Our approach is based on the affinity between keywords and uses the search of cycles in a graph to find the aggregated keywords. We also present performances and a comparison with three other methods. The experimental study shows good results for our approach.