Computing appropriate representations for multidimensional data

On-Line Analytical Processing (OLAP) provides an interactive query-driven analysis of multidimensional data based on a set of navigational operators like roll-up or slice and dice. In most cases, the analyst is expected to use these operations intuitively to find interesting patterns in a huge amount of data of high dimensionality.In this paper, we propose an approach to enhance this analysis by preparing the data set so that the analyst can explore it in a more systematic and effective manner. More precisely we define a measurement of the quality of the representation of multidimensional data and we present a framework for investigating the computation of appropriate representations. We identify the problems of computing such representations and study them w.r.t. an OLAP restructuring operator.