Approximatequeryprocessinghasemergedasanapproachtodealingwiththehugedatavolumeandcomplexqueriesintheenvironmentofdatawarehouse.Inthispaper,wepresentanovelmethodthatprovidesapproximateanswerstoOLAPqueries.Ourmethodisbasedonbuildingacompressed(approximate)datacubebyaclusteringtechniqueandusingthiscompresseddatacubetoprovideanswerstoqueriesdirectly,soitimprovestheperformanceofthequeries.WealsoprovidethealgorithmoftheOLAPqueriesandtheconfidenceintervalsofqueryresults.AnextensiveexperimentalstudywiththeOLAPcouncilbenchmarkshowstheeffectivenessandscalabilityofourcluster-basedapproachcomparedtosampling.
