Estimation of Pollution Solubility in Wastewater by Fusion of Expert Knowledge with Data using the Belief Functions Theory

摘要

In this paper, we propose a methodology for combining expert knowledge with information extracted from statistical data, for estimating pollution solubility in wastewater. The method is based on (1) a case-based approach allowing to predict a quantity of interest from past cases in the form of a belief function, (2) Bayesian networks for modelling expert knowledge and (3) a tuning mechanism allowing to mix information sources, so as to minimize a performance criterion. The use of this method for this environmental problem is motivated by the fact that knowledge in this domain is very partial and ill- structured. The belief functions theory allows to handle the induced uncertainty and imprecision. The approach is expected to be useful in situations where both small databases and partial expert knowledge are available.