Additive Belief-Network Models

The inherent intractability of probabilistic in­ ference has hindered the application of be­ lief networks to large domains. Noisy OR­ gates [30] and probabilistic similarity net­ works [18, 17) escape the complexity of infer­ ence by restricting model expressiveness. Re­ cent work in the application of belief-network models to time-series analysis and forecasting [9, 10) has given rise to the additive belief­ network model (ABNM). We (1) discuss the nature and implications of the approxima­ tions made by an additive decomposition of a belief network, (2) show greater efficiency in the induction of additive models when avail­ able data are scarce, (3) generalize proba­ bilistic inference algorithms to exploit the ad­ ditive decomposition of ABNMs, ( 4) show greater efficiency of inference, and (5) com­ pare results on inference with a simple addi­ tive belief network.