Sensitivity analysis of neural networks to input variation is an important research area as it goes some way to addressing the criticisms of their black-box behaviour. Such analysis of RBFNs for hydrological modelling has previously been limited to exploring perturbations to both inputs and connecting weights. In this paper, the backward chaining rule that has been used for sensitivity analysis of MLPs, is applied to RBFNs and it is shown how such analysis can provide insight into physical relationships. A trigonometric example is first presented to show the effectiveness and accuracy of this approach for first order derivatives alongside a comparison of the results with an equivalent MLP. The paper presents a real-world application in the modelling of river stage shows the importance of such approaches helping to justify and select such models.
References
[1]
Ghorbani, M.A., Zadeh, H.A., Isazadeh, M. and Terzi, O. (2016) A Comparative Study of Artificial Neural Networks (MLP, RBF) and Support Vector Machine Models for River Flow Prediction. Environmental Earth Sciences, 75, 1-14. https://doi.org/10.1007/s12665-015-5096-x
[2]
Nie, S., Bian, J, Wan, H., Sun, X. and Zhang, B. (2017) Simulation and Uncertainty Analysis for Groundwater Levels Using Radial Basis Function Neural Network and Support Vector Machine Models. Journal of Water Supply: Research and Technology, 66, 15-24. https://doi.org/10.2166/aqua.2016.069
[3]
Nawaz, N., Harun, S., Othman, R. and Heryansyah, A. (2016) Application of Radial Basis Function Neural Networks for Modelling Rainfall-Runoff Processes: A Case Study of Semenyih River Catchment, Malaysia. Chiang Mai Journal of Science, 43, 1358-1367.
[4]
Chai, S.S., Wong, W.K. and Goh, K.L. (2017) Rainfall Classification for Flood Prediction Using Meteorology Data of Kuching, Sarawak, Malaysia: Backpropagation vs Radial Basis Function Neural Network. International Journal of Environmental Science and Development, 8, 385-388. https://doi.org/10.18178/ijesd.2017.8.5.982
[5]
Broomhead, D.S. and Lowe, D. (1988) Multivariable Function Interpolation and Adaptive Networks. Complex Systems, 2, 321-355.
[6]
Moody, J.E. and Darkin, C.J. (1989) Fast Learning in Networks of Locally-Tuned Processing Units. Neural Computation, 1, 281-294. https://doi.org/10.1162/neco.1989.1.2.281
[7]
Park, J. and Sandberg, I.W. (1993) Approximation and Radial Basis Function Networks. Neural Computation, 5, 305-316. https://doi.org/10.1162/neco.1993.5.2.305
[8]
Powell, M.J.D. (1987) Radial Basis Functions for Multivariable Interpolation: A Review. In: Mason, J.C. and Cox, M.G., Eds., Algorithms for Approximation, Clarendon Press, Oxford, 143-167.
[9]
Chen, X., Zeng, X., Chu, R. and Zhong, S. (2010) A Quantified Sensitivity Measure of Radial Basis Function Neural Networks to Input Variation. IJCNN, 2010 Joint Conference on Neural Networks, Barcelona, 18-23 July 2010, 1-6. https://doi.org/10.1109/IJCNN.2010.5596949
[10]
Stevenson, M., Winter, R. and Widrow, B. (1990) Sensitivity of Feed forward Neural Networks to Weight Errors. IEEE Transactions on Neural Networks, 1, 71-80. https://doi.org/10.1109/72.80206
[11]
Piche, S.W. (1995) The Selection of Weight Accuracies for Madalines. IEEE Transactions on Neural Networks, 6, 432-445. https://doi.org/10.1109/72.363478
[12]
Zeng, X., Wang, Y. and Zhang, K. (2006) Computation of Adaline’s Sensitivity to Weight Perturbations. IEEE Transactions on Neural Networks, 17, 515-519. https://doi.org/10.1109/TNN.2005.863418
[13]
Hashem, S. (1992) Sensitivity Analysis for Feed forward Artificial Neural Networks with Differentiable Activation Functions. International Joint Conference on Neural Networks 1992, Vol. 1, Baltimore, 419-424.
[14]
Yang, J., Zeng, X.Q. and Zhong, S.M. (2013) Computation of Multilayer Perceptron Sensitivity to Input Perturbation. Neurocomputing, 99, 390-398. https://doi.org/10.1016/j.neucom.2012.07.020
[15]
Zeng, X.Q. and Yeung, D.S. (2001) Sensitivity Analysis of Multilayer Perceptron to Input and Weight Perturbations. IEEE Transactions on Neural Networks, 12, 1358-1366. https://doi.org/10.1109/72.963772
[16]
Huang, L.H., Zeng, X.Q., Zhong, S.M. and Han, L.X. (2014) Sensitivity Study of Binary Feed forward Neural Networks. Neurocomputing, 136, 268-280. https://doi.org/10.1016/j.neucom.2014.01.005
[17]
Ng, W.W.Y., Yeung, D.S., Ran, Q. and Tsang, E.C.C. (2002) Statistical Output Sensitivity to Input and Weight Perturbations of Radial Basis Function Neural Networks. IEEE IC-SMC, Tunisia, 6-9 October 2002, 503-508.
[18]
Shi, D., Yeung, D.S. and Gao, J. (2005) Sensitivity Analysis Applied to the Construction of Radial Basis Function Networks. IEEE Transactions on Neural Networks, 18, 951-957. https://doi.org/10.1016/j.neunet.2005.02.006
[19]
Wang, X., Li, C., Yeung, D.S., Song, S. and Feng, H. (2008) A Definition of Partial Derivative of Random Functions and Its Application to RBFNN Sensitivity Analysis. Neurocomputing, 71, 1515-1526. https://doi.org/10.1016/j.neucom.2007.05.005
[20]
Wing, W.Y., Ng, Z.H., Yeung, D.S. and Chan, P.K. (2014) Steganalysis Classifier Training via Minimizing Sensitivity for Different Imaging Sources. Information Sciences, 281, 211-224. https://doi.org/10.1016/j.ins.2014.05.028
[21]
Harpham, C. and Dawson, C.W. (2006) The Effect of Different Basis Functions on a Radial Basis Function Network for Time Series Prediction: A Comparative Study. Neurocomputing, 69, 2161-2170. https://doi.org/10.1016/j.neucom.2005.07.010
[22]
Dawson, C.W., Abrahart, R.J. and See, L.M. (2007) HydroTest: A Web-Based Toolbox of Evaluation Metrics for the Standardised Assessment of Hydrological Forecasts. Environmental Modelling and Software, 22, 1034-1052. https://doi.org/10.1016/j.envsoft.2006.06.008
[23]
Dawson, C.W., See, L.M., Abrahart, R.J. and Heppenstall, A.J. (2006) Symbiotic Adaptive Neuro-Evolution Applied to Rainfall-Runoff Modelling in Northern England. Neural Networks, 19, 236-247. https://doi.org/10.1016/j.neunet.2006.01.009
[24]
Dawson, C.W., Abrahart, R.J. and See, L.M. (2010) HydroTest: Further Development of a Web Resource for the Standardised Assessment of Hydrological Models. Environmental Modelling & Software, 25, 1481-1482. https://doi.org/10.1016/j.envsoft.2009.01.001
