Publication Abstracts
Aires et al. 2004
Aires, F., C. Prigent, and
, 2004: Neural network uncertainty assessment using Bayesian statistics with application to remote sensing: 3. Network Jacobians. J. Geophys. Res., 109, D10305, doi:10.1029/2003JD004175.Used for regression fitting, Neural Network (NN) models can be used effectively to represent highly nonlinear, multivariate functions. In this situation, most emphasis has been on estimating the output errors but almost no attention has been given to errors associated with the internal structure of the NN model. The complex network of dependency inside the NN is the essence of the model and assessing its quality, coherency, and physical character makes all the difference between a "black-box" with small output errors and a reliable, robust, and physically coherent model. Such dependency structures can, for example, be described by the NN Jacobians: they indicate the sensitivity of one output with respect to the inputs of the model. Estimating these Jacobians is essential for many other applications as well. We use a new method of uncertainty estimate developed in companion paper 1 to investigate the robustness of the quantities that characterize the NN structure. A regularization strategy based on principal component analysis is proposed to suppress the multi-collinearities that are a major concern when analyzing the internal structure of such a model. The theory is applied to the remote sensing application already presented in companion papers 1 and 2.
- Get PDF (360 kB)
- PDF documents require the free Adobe Reader or compatible viewing software to be viewed.
- Go to journal article webpage
Export citation: [ BibTeX ] [ RIS ]
BibTeX Citation
@article{ai02100q, author={Aires, F. and Prigent, C. and Rossow, W. B.}, title={Neural network uncertainty assessment using Bayesian statistics with application to remote sensing: 3. Network Jacobians}, year={2004}, journal={Journal of Geophysical Research}, volume={109}, pages={D10305}, doi={10.1029/2003JD004175}, }
[ Close ]
RIS Citation
TY - JOUR ID - ai02100q AU - Aires, F. AU - Prigent, C. AU - Rossow, W. B. PY - 2004 TI - Neural network uncertainty assessment using Bayesian statistics with application to remote sensing: 3. Network Jacobians JA - J. Geophys. Res. JO - Journal of Geophysical Research VL - 109 SP - D10305 DO - 10.1029/2003JD004175 ER -
[ Close ]