Histograms for OLAP and Data-Stream Queries

Histograms are an important tool for data reduction both in the field of data-stream querying and in OLAP, since they allow us to represent large amount of data in a very compact structure, on which both efficient mining techniques and OLAP queries can be executed. Significant time- and memory-cost advantages may derive from data reduction, but the trade-off with the accuracy has to be managed in order to obtain considerable improvements of the overall capabilities of mining and OLAP tools. In this chapter we focus on histograms, that are shown in the recent literature to be one of the possible concrete answers to the above requirements.