基于案例推理中案例表示的研究  

Researches on Case-Representation in Case-Based Reasoning System

作　　者：李建洋[1] 陈雪云[2] 刘慧婷[1] 谢秀珍[2] 

机构地区：[1]合肥工业大学计算机网络所,合肥230009 [2]龙岩学院计算机科学系,福建龙岩364000

出　　处：《合肥学院学报（自然科学版）》2007年第3期26-29,共4页Journal of Hefei University :Natural Sciences

基　　金：福建省自然科学基金项目(A0640001);福建省教育厅科研基金项目(JA06043);安徽省教育厅科研基金项目(KJ2007B385C)资助

摘　　要：基于案例推理利用过去的案例或经验进行推理来求解新问题,是一种重要的机器学习方法.在CBR(Case-based Reason ing)中知识表示以案例表示为基础,案例是一段带有上下文信息的知识,该知识表达了推理机在达到其目标的过程中能起关键作用的经验.案例表示可能是半结构化或非结构化的,甚至用自然语言来表达的,涉及系统的运行效率.分析了目前的多种表示方法,研究了不同环境下的案例表示,以及选用合适的方法来构筑的案例库.Case-based Reasoning（CBR） as a cognitive model suggests that people learn it the best from storing former problem-solving cases that they solve new problems, and plays an important role in Machine Learning methods. Cases are the contexts that remindings of right cases are fodder for generalization, and present the key experience in CBR process. They may be represented in semistructured or unstructured models, even in natural language text, that is strongly interrelated with the system efficiency. Analyzing the most case-representation methods, the paper presents different strategies to approach case-representation in the status of different conditions, and suggests choosing the most suitable style for the construction of case-base.

关 键 词：基于案例推理 案例表示 案例形式 案例组织 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]