Map Reduce Approach For Computing Interesting Measure For Data Cube
Efficient computation of aggregations plays important part in Data Warehouse systems. Multidimensional data
analysis applications are looking for variations or unusual patterns. They aggregate data across many dimensions. For
aggregation, the SQL aggregate functions and the GROUP BY operator are used. But Data analysis applications need the Ndimensional
generalization of these operators. Data Cube is introduced which is a way of structuring data in N-dimensions so
as to perform analysis over some measure of interest. Data cube computation is a key task in data warehouse. There are
several methods, techniques for cube computation. But these techniques have limitation so new MapReduce based approach
is used. Using Data Partition and Batch formation techniques, data and computation workload is effectively distributed using
MapReduce framework. Extreme data skew is detected. MapReduce based algorithm is used for computing cube in parallel
using partially algebraic measures and will get final Measures by Cube groups aggregations. Interesting cube groups are
identified. Performance of proposed approach is evaluated.
Keywords - Data Cube, Data Cube Computation, Data Cube Mining, MapReduce, Partial algebraic measure.