Towards an Integration of Knowledge in OLAP Cubes Exploratory Analysis

Abstract

This paper presents a general framework allowing a tight coupling between data warehousing and data mining capabilities in a uniform and integrated way. The proposed approach consists in integra­ting knowledge into data warehouses and especially in OLAP cubes exploration. This integration aims to make the OLAP process guided both by multidimensional data and knowledge. A multidimensional structure is then proposed to support data and knowledge. We propose also a set of analytical operators and multidimensional data mining operators. To illustrate the proposed approach, a real OLAP cube built to understand the behaviour of users navigating on the Web is proposed and used to show by concrete examples the interest of such extension of data warehouses.