A study of the bootstrap method for estimating the accuracy of artificial neural networks in predicting nuclear transient processes

The quantification of the uncertainty associated to the results provided by artificial neural networks is essential for their confident and reliable use in practice. This is particularly true for control and safety applications in critical technologies such as those of the nuclear industry. In this paper, the results of a study concerning the use of the bootstrap method for quantifying the uncertainties in the output of supervised neural networks are reported. A thorough parametric analysis is performed with reference to a literature problem. A case study is then provided, concerning the prediction of the feedwater flow rate in a Boiling Water Reactor.

[1]  Halbert White,et al.  Bootstrapping Confidence Intervals for Clinical Input Variable Effects in a Network Trained to Identify the Presence of Acute Myocardial Infarction , 1995, Neural Computation.

[2]  Paolo F. Fantoni,et al.  On-Line Monitoring of Instrument Channel Performance in Nuclear Power Plant Using PEANO , 2002 .

[3]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[4]  David Hinkley,et al.  Bootstrap Methods: Another Look at the Jackknife , 2008 .

[5]  James L. McClelland.  Parallel Distributed Processing , 2005 .

[6]  Da Ruan,et al.  Introduction: Intelligent systems for process monitoring and diagnosis , 2002, Int. J. Intell. Syst..

[7]  Robert E. Uhrig,et al.  A Novel Approach to Process Modeling for Instrument Surveillance and Calibration Verification , 2003 .

[8]  Ron Kohavi,et al.  A Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection , 1995, IJCAI.

[9]  J. Wesley Hines,et al.  Instrument Surveillance and Calibration Verification Through Plant Wide Monitoring Using Autoassociative Neural Networks , 1996 .

[10]  B. Efron.  Bootstrap Methods: Another Look at the Jackknife , 1979 .

[11]  Belle R. Upadhyaya,et al.  Sensor validation for power plants using adaptive backpropagation neural network , 1990 .

[12]  S. Roberts,et al.  1 Confidence Intervals and Prediction Intervals for Feed-Forward Neural Networks a , 2000 .

[13]  Tom Heskes,et al.  Practical Confidence and Prediction Intervals , 1996, NIPS.

[14]  Belle R. Upadhyaya,et al.  Application of Neural Networks for Sensor Validation and Plant Monitoring , 1990 .

[15]  Da Ruan,et al.  Power Plant Surveillance and Diagnostics , 2002 .

[16]  S. Roberts,et al.  Confidence Intervals and Prediction Intervals for Feed-Forward Neural Networks , 2001 .

[17]  H. M. Hashemian,et al.  On-line testing of calibration of process instrumentation channels in nuclear power plants. Phase 2, Final report , 1995 .

[18]  Elie Bienenstock,et al.  Neural Networks and the Bias/Variance Dilemma , 1992, Neural Computation.

[19]  Léon Personnaz,et al.  CONSTRUCTION OF CONFIDENCE INTERVALS IN NEURAL MODELING USING A LINEAR TAYLOR EXPANSION , 1998 .

[20]  Robert E. Uhrig.  Regulatory Treatment of On-Line Surveillance and Diagnostic Systems , 2002 .

[21]  Davide Roverso.  Fault diagnosis with the Aladdin transient classifier , 2003, SPIE Defense + Commercial Sensing.

[22]  J. Wesley Hines,et al.  IMPLEMENTATION OF ON-LINE MONITORING PROGRAMS AT NUCLEAR POWER PLANTS , 2004 .

[23]  Robert Tibshirani,et al.  A Comparison of Some Error Estimates for Neural Network Models , 1996, Neural Computation.

[24]  Robert Tibshirani,et al.  An Introduction to the Bootstrap , 1994 .