|
|
|
|
Error Estimation in Approximate Bayesian Belief Network Inference
Enrique Castillo, Remco Bouckaert, Jose Sarabia, Cristina Solares
Abstract:
We can perform inference in Bayesian belief networks by enumerating instantiations with high probability thus approximating the marginals. In this paper, we present a method for determining the fraction of instantiations that has to be considered such that the absolute error in the marginals does not exceed a predefined value. The method is based on extreme value theory. Essentially, the proposed method uses the reversed generalized Pareto distribution to model probabilities of instantiations below a given threshold. Based on this distribution, an estimate of the maximal absolute error if instantiations with probability smaller than u are disregarded can be made.
Keywords: Approximate inference, Bayesian networks, extremes, error bounds.
Pages: 55-62
PS Link:
PDF Link: /papers/95/p55-castillo.pdf
BibTex:
@INPROCEEDINGS{Castillo95,
AUTHOR = "Enrique Castillo
and Remco Bouckaert and Jose Sarabia and Cristina Solares",
TITLE = "Error Estimation in Approximate Bayesian Belief Network Inference",
BOOKTITLE = "Proceedings of the Eleventh Conference Annual Conference on Uncertainty in Artificial Intelligence (UAI-95)",
PUBLISHER = "Morgan Kaufmann",
ADDRESS = "San Francisco, CA",
YEAR = "1995",
PAGES = "55--62"
}
|
|
