Data Mining Techniques in Simulation Results Analysis

Abstract

The research is carried out in the area of analysis of simulation results by using data mining techniques. The goal of this research is to explore the applicability of data mining techniques in the area of simulation result analysis, offer data mining techniques application scheme in analysis of simulation results, as well as demonstrate the usage of these techniques in analysis of experimental data. As a result of the theoretical study, an approach is proposed, consisting of two stages and combining the fundamental principles of data farming and knowledge discovery. A variety of data mining techniques, including correlation analysis, clustering and several visualization mechanisms of results are used for the knowledge discovery. The approach is applied to the analysis of experimental data. The performence of a queueing system was analyzed, and knowledge and decision rules are obtained from simulation results.