Proper and effective training of a pattern recognizer for cyclic data

A new approach to training backpropagation neural networks for identifying cyclic patterns on control charts is presented. The objectives of this research are to show that building an effective cyclic-pattern-recognition neural network requires proper training strategies and to demonstrate how these strategies, namely, incremental and decremental training, should be applied and how the performance can be improved with additional statistics. A series of experiments were conducted to study the effect of the number of output pattern classes and the effect of noise on network training and performance. Experiments show that reducing the number of output pattern classes to a small number, e.g., four or fewer, does not guarantee effective learning, and that the noise added to the training data should be maintained at a reasonable level to achieve a balanced performance. Further incorporation of harmonic amplitude statistics (HAS) also proved that the proper use of statistics adopted from Fourier analysis can imp...

[1]  C. Chatfield,et al.  Fourier Analysis of Time Series: An Introduction , 1977, IEEE Transactions on Systems, Man, and Cybernetics.

[2]  Geoffrey E. Hinton,et al.  Learning internal representations by error propagation , 1986 .

[3]  H. B. Hwarng A study of the learning of three-layer back-propagation pattern recognizers for statistical quality control charts , 1991, IJCNN-91-Seattle International Joint Conference on Neural Networks.

[4]  N F Hubele,et al.  X¯ CONTROL CHART PATTERN IDENTIFICATION THROUGH EFFICIENT OFF-LINE NEURAL NETWORK TRAINING , 1993 .

[5]  Alice E. Smith,et al.  X-bar and R control chart interpretation using neural computing , 1994 .

[6]  Kevin J. Dooley,et al.  Identification of change structure in statistical process control , 1992 .

[7]  B. L. Foote,et al.  Spectral analysis in quality control: a control chart based on the periodogram , 1988 .

[8]  B. H. Gwee,et al.  Cause associator network for fuzzily deduced conclusion in process control , 1991, [Proceedings] 1991 IEEE International Joint Conference on Neural Networks.

[9]  Norma Faris Hubele,et al.  Back-propagation pattern recognizers for X¯ control charts: methodology and performance , 1993 .

[10]  H. B. Hwarng Multilayer perceptions for detecting cyclic data on control charts , 1995 .

[11]  J. Spurrier,et al.  Control charts for detecting cyclical behavior , 1990 .

[12]  Herman Sahrmann,et al.  Set-Up Assurance Through Time Series Analysis , 1979 .

[13]  H. Brian Hwarng,et al.  Boltzmann machines that learn to recognize patterns on control charts , 1992 .