Approximate Selection Queries over Imprecise Data

• Authors: Iosif Lazaridis;Sharad Mehrotra
• Venue: ICDE '04 Proceedings of the 20th International Conference on Data Engineering
• Year: 2004

We examine the problem of evaluating selection queriesover imprecisely represented objects. Such objects are usedeither because they are much smaller in size than the preciseones (e.g., compressed versions of time series), or asimprecise replicas of fast-changing objects across the network(e.g., interval approximations for time-varying sensorreadings). It may be impossible to determine whether an impreciseobject meets the selection predicate. Additionally,the objects appearing in the output are also imprecise. Retrievingthe precise objects themselves (at additional cost)can be used to increase the quality of the reported answer.In our paper we allow queries to specify their own answerquality requirements. We show how the query evaluationsystem may do the minimal amount of work to meetthese requirements. Our work presents two important contributions:first, by considering queries with set-based answers,rather than the approximate aggregate queries overnumerical data examined in the literature; second, by aimingto minimize the combined cost of both data processingand probe operations in a single framework. Thus, we establishthat the answer accuracy/performance tradeoff canbe realized in a more general setting than previously seen.