Ngo L, Haddawy P, Krieger R A, Helwig J
Department of Electrical Engineering and Computer Science, University of Wisconsin-Milwaukee 53201, USA.
Comput Biol Med. 1997 Sep;27(5):453-76. doi: 10.1016/s0010-4825(97)00015-2.
We present a language for representing context-sensitive temporal probabilistic knowledge. Context constraints allow inference to be focused on only the relevant portions of the probabilistic knowledge. We provide a declarative semantics for our language. We present a sound and complete algorithm for computing posterior probabilities of temporal queries, as well as an efficient implementation of the algorithm. Throughout we illustrate the approach with the problem of reasoning about the effects of medications and interventions on the state of a patient in cardiac arrest. We empirically evaluate the efficiency of our system by comparing its inference times on problems in this domain with those of standard Bayesian network representations of the problems.
我们提出一种用于表示上下文相关的时态概率知识的语言。上下文约束使得推理能够仅聚焦于概率知识的相关部分。我们为我们的语言提供了一种声明式语义。我们提出一种用于计算时态查询后验概率的合理且完备的算法，以及该算法的高效实现。在整个过程中，我们用关于药物和干预措施对心脏骤停患者状态影响的推理问题来说明该方法。我们通过将我们的系统在该领域问题上的推理时间与这些问题的标准贝叶斯网络表示的推理时间进行比较，对我们系统的效率进行了实证评估。
