Statistical analysis in EMC using dimension reduction methods

Many proposed efficient statistical analysis methods in EMC are limited due to the dimensionality problem; when the number of random variables becomes large the methods can become less efficient than using the established Monte Carlo method. In this paper the univariate and bivariate dimension reduction methods are examined for their applicability and efficiency for statistical EMC analysis. The performance of the techniques is evaluated using an example of coupling between wires within an enclosure and compared with the Monte Carlo Method.