Probabilistic asthma case finding: a noisy or reformulation.

摘要: Bayesian Networks are used to model domain knowledge with natural perception of causal influences. Even though reduce the number probabilities required specify relationships in domain, specifying these for large networks can be prohibitive. The Noisy-OR formalism Network (BN) overcomes this shortcoming by making an assumption independence among modeled causes and their common effect. However, accuracy has rarely been tested. In paper we report results empirical study asthma case finding that compares reformulation expert BN trained using clinical data set from Regenstrief Medical Record System. Our show formulation performs comparably expert’s suggesting is robust, at least domain.