Management of context-aware preferences in multidimensional databases

Abstract

Personalization of information systems brings new challenges to OLAP technology. A key characteristic of emerging OLAP database systems will be customizability of their behaviour taking into account users preferences as well as their context of analysis. In this paper we define a personalization framework for OLAP database systems based on user context-aware preferences. We consider a qualitative preference model which handles user preferences on the multidimensional schema. Context is modelled as a tree of multidimensional components of an OLAP analysis (fact, measures, dimensions, parameters). We define some OLAP operations that support personalization. User queries are dynamically enhanced with his preferences and are aware of the analysis context.