A neural network learning method for causal networks
Yun Peng · 2002
This paper presents a neural network method that learns both symbolic and probabilistic causal associations for probabilistic causal networks. Unlike past neural network modeling work, this method directly acts on causal networks without requiring their own separate networks, and it learns either from a set of static case data or upon receiving a new case input. Theoretical analyses and computer experiments of this method are also presented.>