Asymptotic bias correction for a probabilistic neural network for structural health monitoring

摘要

This paper describes recent work on minimizing decision errors in a probabilisitic neural network used for change point detection in structures. The relation between bias error in the kernel density estimator and the training set size is examined.