Theory Refinement on Bayesian Networks
This work addresses theory refinement for domain experts in Bayesian statistics, but it is incremental as it adapts existing batch algorithms to incremental learning without introducing new paradigms.
The paper tackles the problem of automatically updating domain theories under uncertainty, framing it as an incremental learning task where a system starts with a partial expert-provided theory and refines it from data. It presents algorithms for refining Bayesian networks, which are incremental variants of existing batch methods, enabling effective operation in both batch and incremental modes.
Theory refinement is the task of updating a domain theory in the light of new cases, to be done automatically or with some expert assistance. The problem of theory refinement under uncertainty is reviewed here in the context of Bayesian statistics, a theory of belief revision. The problem is reduced to an incremental learning task as follows: the learning system is initially primed with a partial theory supplied by a domain expert, and thereafter maintains its own internal representation of alternative theories which is able to be interrogated by the domain expert and able to be incrementally refined from data. Algorithms for refinement of Bayesian networks are presented to illustrate what is meant by "partial theory", "alternative theory representation", etc. The algorithms are an incremental variant of batch learning algorithms from the literature so can work well in batch and incremental mode.