Pre-aggregation for Probability Distributions

Igor Timko, Curtis E. Dyreson, Torben Bach Pedersen

Publikation: Bog/antologi/afhandling/rapportRapportForskning

Abstrakt

Motivated by the increasing need to analyze complex uncertain multidimensional data (e.g., in order to optimize and personalize location-based services), this paper proposes novel types of {\em probabilistic} OLAP queries that operate on aggregate values that are probability distributions and the techniques to process these queries. The paper also presents the methods for computing the probability distributions, which enables pre-aggregation, and for using the pre-aggregated distributions for further aggregation. In order to achieve good time and space efficiency, the methods perform approximate computations of aggregate values. The paper also reports on the experiments with the methods. The work is motivated with a real-world case study, based on our collaboration with a leading Danish vendor of location-based services. No previous work considers the combination of the aspects of uncertain multidimensional data analysis that is considered in this paper (i.e., approximate processing of probabilistic OLAP queries over probability distributions).
OriginalsprogEngelsk
Udgivelses stedAalborg
ForlagDepartment of Computer Science, Aalborg University
Antal sider56
StatusUdgivet - 2005
NavnDB Tech Report
Nummer14

Fingeraftryk Dyk ned i forskningsemnerne om 'Pre-aggregation for Probability Distributions'. Sammen danner de et unikt fingeraftryk.

  • Citationsformater

    Timko, I., Dyreson, C. E., & Pedersen, T. B. (2005). Pre-aggregation for Probability Distributions. Department of Computer Science, Aalborg University. DB Tech Report, Nr. 14