A Parallel distributed Cluster Computing Model for Z-Score Computation in Respect of Sri Lankan University Admissions

The Z-score or Standard Score, a numerical measurement used in statistics, is calculated based on the mean value and standard deviation of groups. In Sri Lanka, the Z- score is being used for evaluating advanced level students’ performance based on their subject wise results. It is further used to rank students district wise and island wide. As number of students are increasing in each year for advanced level examination, computation of z-score needs more computing resources. Computation of Z-score on a large dataset is a resource-intensive task. To solve this computation process, it is essential to address parallelising operations to speed up the computation and sustain less total processing power when it is distributed to tiny processing nodes. The present research study investigates and experiments the feasibility of parallelizing the z-score computation using existing computing resources connected through the local network as clusters, without affecting the existing infrastructure. The parallel cluster model reveals the significant improvement with traditional sequential method in terms of execution time and speedup for the computation.