SVD-NET: an algorithm that automatically selects network structure

  • Authors:
  • D. C. Psichogios;L. H. Ungar

  • Affiliations:
  • Dept. of Chem. Eng., Pennsylvania Univ., Philadelphia, PA;-

  • Venue:
  • IEEE Transactions on Neural Networks
  • Year:
  • 1994

Quantified Score

Hi-index 0.00

Visualization

Abstract

An algorithm is developed for training feedforward neural networks that uses singular value decomposition (SVD) to identify and eliminate redundant hidden nodes. Minimizing redundancy gives smaller networks, producing models that generalize better and thus eliminate the need of using cross-validation to avoid overfitting. The method is demonstrated by modeling a chemical reactor