Domain Partitioning: Approach to Computing Average Iceberg Queries

Abstract

: Data analytics and data mining systems work on data which stored in files, the files are not store relationships among the data, from such kind of data we compute aggregate values over the set of required attributes for find insights of data, find attributes values which aggregation values greater than threshold such kind of queries called iceberg queries. Computing iceberg queries with average aggregate function is default, because limited memory available. Existing method are suffers with re-computation of candidate. We proposed a Record Traction Algorithm(RTA) ,it use Domain partitioning approach, it avoid re-computation of candidate in during next scan of data set, it use bit vector and bitmap numbers for Domain Partitioning the data, our experiment reveals that our approach generate a candidate only once and input data will reduced in further candidate sets.