Using Statistical Analysis To Improve Data Partitioning In Algorithms For Data Parallel Processing Implementation