论文部分内容阅读
所有法律案件均涉及用证据推理。随着软件支持工具的发展,证据推理需要一个形式基础。文献表明:证据推理有三种主要方法,即论辩方法、叙事方法和概率推理。本文组合了后两种方法。贝叶斯网络在法律案件中应用的最新研究表明,当人们用法律贝叶斯网络来再现结构时已经产生了许多法律习语。贝叶斯网络是用来量化案件中各种变量如何进行互动的。在叙事方法中,情节提出了案件的证据语境,但当前缺乏一种方法把定量的贝叶斯网络数值技术与定性的情节合并起来。本文给出了一种用单个贝叶斯网络为几个情节建模。这种方法已经通过了案件研究的检验。我们需要引入两个新的法律习语:一是情节习语,二是合并情节习语。合成网络目的是帮助法官或陪审团的,帮助维护案件中两个相关变量间互动的优异概观,通过比较各种情节来防止视野过于狭窄。
All legal cases involve reasoning with evidence. As software support tools evolve, evidence reasoning needs a formal basis. The literature shows that there are three main methods of evidence reasoning: argumentative method, narrative method and probabilistic reasoning. This article combines the latter two methods. The latest research on the application of Bayesian networks in legal cases shows that many legal idioms have been produced when people use the legal Bayesian network to reproduce their structures. Bayesian networks are used to quantify how variables in a case interact. In the narrative approach, the plot raises the context of the evidence in the case, but there is a lack of a way to combine quantitative Bayesian network numerical techniques with qualitative episodes. This paper presents a few Bayesian network modeling for several episodes. This method has been tested in case studies. We need to introduce two new legal idioms: one is the episodic idiom and the second is the merger of the episodic idioms. The purpose of the synthesis network is to help the judge or jury to help preserve the excellent overview of the interaction between two related variables in the case and to prevent the view from being too narrow by comparing the various scenarios.