Verification of the new method of model selection as objective, by using adaptive-kernel-density-estimator and simulated-ideal-data, in bootstrap-cross-validation for expectation of Bayesian evidence