定罪量刑系统DZLX中不确定性知识的表示和处理

Representation and Management of Uncertain Knowledge in DZLX System

文章首先概述了国内外不确定性知识的处理方法,又通过分析刑法适用中的不确定性问题,介绍了我们开发的“定罪量刑参考系统”DZLX中不确定性知识的表达,推理和求解的方法,由此提出了在不确定性知识处理领域中带有普遍意义的原则,并且对不确定性知识学习方法的机器实现作了探索。

A brief review of the management of uncertain knowledge in expert systems has been given in this paper firstly. Then, after giving a simple analysis of the puzzles about uncertainty in deciding criminal cases, the paper introduces the methodology of the representation, reasoning and problem solving about the uncertain knowledge in the DZLX (short for 'Ding Zui Liang Xing' in Chinese phoneticism), which is used for conviction and penal discretion according to Criminal Law in a court procedure. As a conclusion, the paper presents some general principles for the management of uncertainties in expert systems. The exploration of the machine implementation method for learning uncertain knowledge was also given.