| 摘要: |
| 近年来,已提出了一些非确定性信息的理论,如概率理论,证据理论,可能性理论,基于Bayesian公式的似然推理模型等。这些方法均在不同程度上要求专家给出一些先验概率值,从而影响了它们的应用程度。所以,建立一种有效的获取先验概率值的方法是有必要的。本文给出一种从实际例子中自动获取先验概率值的方法。 |
| 关键词: 概率推理 不确定性推理 模糊推理 信度网络 |
| DOI： |
| 投稿时间：1997-06-05 |
| 基金项目:得到国家自然科学基金和国家863计划的资助。 |
|
| A Method of Learning Prior Probability |
|
Chen Jun, Zhang Shichao
|
| (Guangxi Training School of Public Finance, Mingxiulu, Nanning, Guangxi, 530003) |
| Abstract: |
| In recent years, some theory of handling uncertainty information,such as probability theory, evidence theory,possibility theory and plausible reasoning model, are established. Because the methods are all based on prior probability, their applicability is constrained. Hence,it is necessary to establish an efficiency way of acquiring prior probability. An automatic approach that the prior probabilites are acquired from examples is suggested in this paper. |
| Key words: Probabilistic reasoning uncertainty reasoning fuzzy reasoning belief network |
