ML Anthology

A Comparison of Some Error Estimates for Neural Network Models

Robert Tibshirani
NeCo 1996 pp. 152-163
doi:10.1162/NECO.1996.8.1.152 /neco/1996/tibshirani1996neco-comparison/

Abstract

We discuss a number of methods for estimating the standard error of predicted values from a multilayer perceptron. These methods include the delta method based on the Hessian, bootstrap estimators, and the “sandwich” estimator. The methods are described and compared in a number of examples. We find that the bootstrap methods perform best, partly because they capture variability due to the choice of starting weights.

NeCo Semantic Scholar

Cite

Text

Tibshirani. "A Comparison of Some Error Estimates for Neural Network Models." Neural Computation, 1996. doi:10.1162/NECO.1996.8.1.152

Markdown

[Tibshirani. "A Comparison of Some Error Estimates for Neural Network Models." Neural Computation, 1996.](https://mlanthology.org/neco/1996/tibshirani1996neco-comparison/) doi:10.1162/NECO.1996.8.1.152

BibTeX

@article{tibshirani1996neco-comparison,
  title     = {{A Comparison of Some Error Estimates for Neural Network Models}},
  author    = {Tibshirani, Robert},
  journal   = {Neural Computation},
  year      = {1996},
  pages     = {152-163},
  doi       = {10.1162/NECO.1996.8.1.152},
  volume    = {8},
  url       = {https://mlanthology.org/neco/1996/tibshirani1996neco-comparison/}
}