✦ LIBER ✦

On the generalization error by a layered statistical model with Bayesian estimation

✍ Scribed by Sumio Watanabe

Publisher: John Wiley and Sons
Year: 2000
Tongue: English
Weight: 287 KB
Volume: 83
Category: Article
ISSN: 1042-0967
DOI: 10.1002/(sici)1520-6440(200006)83:6<95::aid-ecjc11>3.0.co;2-b

No coin nor oath required. For personal study only.

✦ Synopsis

When a statistical estimation is carried out with a model that has layered parameters such as a neural network, the behavior of the generalization error and the optimum model design method are unknown, since unlike the regular model, the asymptotic behavior of the estimated parameters is not clear. In this paper, it is theoretically proven that the generalization error can be decreased if the Bayesian method is applied to a layered model. It is shown that the Bayesian method is effective as a learning method for a large-scale layered model.

📜 SIMILAR VOLUMES

Comment: on the estimation of simultaneo

Comment: on the estimation of simultaneous-equations error-components models with an application to a model of developing country foreign trade

✍ Steven S. Vickner; Stephen P. Davies 📂 Article 📅 1998 🏛 John Wiley and Sons 🌐 English ⚖ 49 KB 👁 2 views

## REFERENCE Kinal, T. and K. Lahiri (1993) , `On the estimation of simultaneous-equations error-components models with an application to a model of developing country foreign trade', Journal of Applied Econometrics, 8, 81±92.

A Bayesian analysis of regression models

A Bayesian analysis of regression models with continuous errors with application to longitudinal studies by L. D. Broemeling and P. Cook, Statistics in Medicine, 16, 321–332 (1997)

✍ J. A. Nelder 📂 Article 📅 1998 🏛 John Wiley and Sons 🌐 English ⚖ 83 KB 👁 2 views

their results. As we have discussed with the authors, we believe that there remains much useful work to be done to evaluate the estimator and to determine how and where it should be use (Schoenfeld, personal communication). We also think that, in the absence of censoring, situations where the new te