Dense sample data are acquired by using numerical control platform of high precision, and the Bayesian generalization is adopted during training the neural network.

• 通过高精度的数控移动工件台获取密集的样本数据，并在神经网络训练过程中采用贝叶斯正则化方法。