摘要
贝叶斯网络是一种有效的不确定性知识表达和推理工具,概率推理是其重要研究内容之一。经过二十年的发展,贝叶斯网络已经有一些比较有效的精确和近似推理算法。对迄今为止的贝叶斯网络推理算法研究进行综述,从复杂度、适用性、精度等方面对它们进行比较分析,指出每种算法的关键环节,为实际应用中算法选择和研究提供参考。
Bayesian network (BN) is a powerful tool to express and infer uncertain knowledge. Probabilistic inference is an important aspect of its research. Bayesian networks have already had some relatively mature accurate and approximate inference algorithms as a result of twenty years' development. The present achievement on Bayesian network inference algorithms is surveied. And then a thorough analysis of the algorithms'complexity, applicability and precision is presented. Some key aspects of the algorithms are also pointed out. The survey will be helpful on selection and research of the inference algorithms.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2008年第5期935-939,共5页
Systems Engineering and Electronics
关键词
贝叶斯网络
精确推理
近似推理
Bayesian network
accurate inference
approximate inference