The use of Tikhonov smoothing functional for solution of the problem of reconstruction of multidimensional nonlinear functions in multilayer perceptron (MLP - networks)

Return to list of articles
Doklady Bashkirskogo Universiteta. 2017. Volume 2. No. 3. pp. 413-420.

Authors

Biryukov A. N.*
Bashkir State University, Sterlitamak Branch
49 Lenin Street, 453103 Sterlitamak, Republic of Bashkortostan, Russia

Abstract

This article discusses the issue of implementation of conceptual basis in specific methods for solving inverse problems of interpretation and the methods of regularization using the principles of Bayesian and Tikhonov. Developed a method to assess the adequacy of neural network models in the absence of any a priori information about the distribution law of noise in the database.

Keywords

  • neural network model (NSM)
  • MLP network
  • the Tikhonov parametric functional
  • Bayesian regularization
  • cross-validation
Return to list of articles