Theory Refinement on Bayesian Networks

Theory refinement is the task of updating a domain theory in the light of new cases, to be done automatically or with some expert as­ sistance. The problem of theory refinement under uncertainty is reviewed here in the con­ text of Bayesian statistics, a theory of belief revision. The problem is reduced to an incre­ mental learning task as follows: the learning system is initially primed with a partial the­ ory supplied by a domain expert, and there­ after maintains its own internal representa­ tion of alternative theories which is able to be interrogated by the domain expert and able to be incrementally refined from data. Algo­ rithms for refinement of Bayesian networks are presented to illustrate what is meant by "partial theory", "alternative theory repre­ sentation", etc. The algorithms are an incre­ mental variant of batch learning algorithms from the literature so can work well in batch and incremental mode.