Personalization of OLAP queries

Personalizing e-services by allowing users to expre ss preferences is becoming more and more common. When querying, expressing preferences is se en as a natural way to avoid empty results or information flooding as well as a solution reducing the user effort in carrying out complex searches. This is particularly beneficial in the OLAP context , since multidimensional databases store huge amounts of data. Besides, OLAP queries have a compl ex structure: they do not only formulate selections and projections on attributes and measur s, they also specify on what hierarchical attributes data are to be aggregated. Finally, duri ng an OLAP session, the user often does not exactly know what she is looking for. The reasons b ehind a specific phenomenon or trend may be hidden, and finding those reasons by manually apply ing different combinations of OLAP operators may be very frustrating.