Wavelet Neural Network-based Control Chart Patterns Recognition

As for six patterns of control chart, wavelet neural network was presented. The structure of wavelet neural network was made up of three-layer network, of which there were 24 neurons of input layer, 20 neurons of hidden layer as well as 6 neurons of output layer. Mexican Hat wavelet was selected as activation function. The simulation results show that wavelet neural network have many advantages, such as quicker training and better recognition performance and better convergence than BP networks, which can be used in control chart recognition.