Model selection in neural networks: Some difficulties

Model selection in neural networks: Some difficulties

0.00 Avg rating0 Votes
Article ID: iaor2007349
Country: Netherlands
Volume: 170
Issue: 2
Start Page Number: 567
End Page Number: 577
Publication Date: Apr 2006
Journal: European Journal of Operational Research
Authors: ,
Keywords: neural networks
Abstract:

This paper considers two related issues regarding feedforward Neural Networks (NNs). The first involves the question of whether the network weights corresponding to the best fitting network are unique. Our empirical tests suggest an answer in the negative, whether using standard Backpropagation algorithm or our preferred direct (non-gradient-based) search procedure. We also offer a theoretical analysis which suggests that there will almost inevitably be functional relationships between network weights. The second issue concerns the use of standard statistical approaches to testing the significance of weights or groups of weights. Treating feedforward NNs as an interesting way to carry out nonlinear regression suggests that statistical tests should be employed. According to our results, however, statistical tests can in practice be indeterminate. It is rather difficult to choose either the number of hidden layers or the number of nodes on this basis.

Reviews

Required fields are marked *. Your email address will not be published.