The concept of the variance estimation for the neural network approximator by jackknife subsampling
Languages of publication
The estimation of a variance for a semi-parametric neural network model variance for geometric properties of sintered metal will be done on the basis of jackknife subsampling method. Calculation results are of great practical significance because it will be possible to use proposed approach in similar microscale modelling. The proposed approach is simple and has many advantages if model identification procedure is computational expensive.
Publication order reference