Bootstrap Model Selection Criterion for Determining the Number of Hidden Units in Neural Network Model

  • Hwang, Changha (Dept. of Statistical Information, Catholic University of Taegu-Hyosung) ;
  • Kim, Daehak (Dept. of Statistical Information, Catholic University of Taegu-Hyosung)
  • 발행 : 1997.12.01

초록

Statistical relations between a system and empirical distribution are studied in terms of the concept of Akaike's Information Criterion. From this consideration we derive a bootstrap criterion for determining the optimal number of hidden units in nerual networks.

키워드

참고문헌

  1. IEEE Transactions on Electronic Computers v.EC-16 Theory of adaptive pattern classifiers Amari, S.
  2. Neurocomputing v.5 Backpropagation and stochastic gradient descent method Amari, S.
  3. Statistics and Probability Letters v.26 A note on bootstrap model selection criterion Chung, H.;Lee, K.;Koo, J.
  4. Model Selection Linhart, H.;Zucchini, W.
  5. Neural Networks v.8 Unbiased estimate of generalization error and model selection in neural network Liu, Y.
  6. Artificial Neural Networks A criterion for determining the number of parameters in an artificial neural network model Murata, N.;Yoshizawa, S.;Amari, S.;T. Kohonen(et al.)(ed.)
  7. IEEE Transactions on Neural Networks v.5 Network information criterion - Determining the number of hidden units for an artificial neural network model Murata, N.;Yoshizawa, S.;Amari, S.