Abstract
The leave-one-out cross-validation scheme for generalization assessment of neural network models is computationally expensive due to replicated training sessions. In this paper we suggest linear unlearning of examples as an approach to approximative cross-validation. Further, we discuss the possibility of exploiting the ensemble of networks offered by leave-one-out for performing ensemble predictions. We show that the generalization performance of the equally weighted ensemble predictor is identical to that of the network trained on the whole training set.
Numerical experiments on the sunspot time series prediction benchmark demonstrate the potential of the linear unlearning technique.
Article PDF
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Avoid common mistakes on your manuscript.
References
H. Akaike, Fitting autoregressive models for prediction, Annals of the Institute of Statistical Mathematics 21, 1969, 243–247.
W.L. Buntine and A.S. Weigend, Bayesian back-propagation, Complex Systems 5, 1991, 603–643.
T. Fox, D. Hinkley and K. Larntz, Jackknifing in nonlinear regression, Technometrics 22, 1980, 29–33.
A. Krogh and J. Vedelsby, Neural network ensembles, cross validation, and active learning, in:Advances in Neural Information Processing Systems 7, eds. G. Tesauro et al., MIT Press, Cambridge, Massachusetts, 1995.
J. Larsen and L.K. Hansen, Generalization performance of regularized neural network models, in:Proceedings of the IEEE Workshop on Neural Networks for Signal Processing IV, eds. J. Vlontzos, J.-N. Hwang and E. Wilson, IEEE, Piscataway, NJ, 1994, pp. 42–51.
J. Larsen and L.K. Hansen, Empirical generalization assessment of neural network models, in:Proceedings of the IEEE Workshop on Neural Networks for Signal Processing V, eds. F. Girosi, J. Makhoul, E. Manolakos and E. Wilson, IEEE, Piscataway, NJ, 1995, pp. 30–39.
L. Ljung,System Identification: Theory for the User, Prentice-Hall, Englewood Cliffs, NJ, 1987.
J. Moody, Note on generalization, regularization, and architecture selection in nonlinear learning systems, in:Proceedings of the 1st IEEE Workshop on Neural Networks for Signal Processing, eds. B.H. Juang, S.Y. Kung and C.A. Kamm, IEEE, Piscataway, NJ, 1991, pp. 1–10.
J. Moody, Prediction risk and architecture selection for neural networks, in:From Statistics to Neural Networks: Theory and Pattern Recognition Applications, eds. V. Cherkassky, J.H. Friedman and H. Wechsler, Series F, vol. 136, Springer, Berlin, 1994.
N. Murata, S. Yoshizawaand and S. Amari, Network information criterion — determining the number of hidden units for an artificial neural network model, IEEE Transactions on Neural Networks, 5, 1994, 865–872.
G.A.F. Seber and C.J. Wild,Nonlinear Regression, Wiley, New York, 1989.
M. Stone, Cross-validatory choice and assessment of statistical predictors, Journal of the Royal Statistical Society B 36(2), 1974, 111–147.
G.T. Toussaint, Bibliography on estimation of misclassification, IEEE Transactions on Information Theory 20(4), 1974, 472–479.
G. Wahba, Spline models for observational data,CBMS-NSF Regional Conference Series in Applied Mathematics, vol. 59, SIAM, 1990.
A.S. Weigend, B.A. Hubermann and D.E. Rumelhart, Predicting the future: a connectionist approach, International Journal of Neural Systems 1(3), 1990, 193–209.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Hansen, L.K., Larsen, J. Linear unlearning for cross-validation. Adv Comput Math 5, 269–280 (1996). https://doi.org/10.1007/BF02124747
Issue Date:
DOI: https://doi.org/10.1007/BF02124747