Application of principal component analysis for optimization of a system operation assessment criteria set

The analysis of results of experimental tests and a literature survey of the issue reveal that the subject matter connected with determination of the optimal number of a given transportation system operation assessment criteria has a direct influence on the result of the considered assessment. The analysis was made within the research on operation quality of a selected transportation system. In order to optimize the analysed system assessment criteria, a theoretical description has been made and an example of canonical correlations application has been presented. Analysis of canonical correlations involves determination of linear combination parameters of the studied sets so that the obtained correlation coefficient will have maximal value. In the successive step, the next pair of linear combination are not correlated with the combinations determined in the first step. The determined correlations can measure the power of relation between two sets of variables and are useful in the process of choosing significant criteria for a given research object operation quality assessment.