9th WSEAS International Conference on Simulation, Modelling and Optimization, Budapest, Hungary, 3 - 05 September 2009, pp.266-267
In this paper we present the theoretical background for destructive component identification in ensemble method via multivariate decompositions. The identification method is based on second order statistics and it is addressed for predictive models scored by MSE criterion. The validity of the concept is confirmed by simulation study based on Friedman benchmark function.