Parameter estimation in directed evidential networks from evidential databases
Narjes Ben Hariz, Boutheina Ben Yaghlane · 2014
Evidential networks are considered as a powerful and flexible tools, commonly used for analyzing complex systems and handling different types of uncertainty in data. A crucial step to benefit from the reasoning process in these models is to quantify them. Thus, we address, in this paper, the issue of estimating parameters in evidential networks from evidential databases, by applying the maximum likelihood estimation generalized to the evidence theory framework.