multilayer artificial neural network performance evaluation apparatus including multiple sub models and method therefor

摘要

In the artificial neural network modeled based on the physical causal relationship, it is possible to evaluate the performance of the artificial neural network according to the combination of the input variable items and sub-models of each sub-model, and accordingly, provide information on the optimal artificial neural network configuration. It relates to a multilayer artificial neural network performance evaluation apparatus including a plurality of sub-models and a method thereof, wherein each input variable input to a plurality of sub-models is increased or decreased, and an evaluation index is calculated by measuring the output of each sub-model according to the increase or decrease. Its purpose is to evaluate the performance of a multilayer artificial neural network including a plurality of sub-models, and to provide design information of an artificial neural network structure showing optimal reliability according to the evaluation result.