摘要
范例推理作为基于规则推理技术的一个重要补充,已受到人工智能研究人员的普遍关注.在范例推理中,也有许多相应的知识,它主要包括:范例库、修正知识库、索引模式、相似性判断标准等.这些知识的获取也存在一定的瓶颈问题.通过在范例推理中引入数据挖掘技术,可望部分解决其知识获取的自动化问题,提高智能系统的整体性能.本文着重讨论了在范例推理中使用数据挖掘技术的必要性,提出了二个自动获取范例库的思想与算法,并应用于一个实际的系统之中.实验结果表明,这种思路大大提高了智能系统的开发周期与系统的运行效率.
As an important supplementary technique to rule based reasoning,case based reasoning, has been emphasized by researchers in the artificial intelligence field. In a case based system, there are many kinds of knowledge, such as case base ,adaptation knowledge base ,indexing model, similarity assessing criteria, etc. There exists also a bottleneck in these knowledge acquisitions. The use of data mining may automat the acquisition of the knowledge and heighten the whole competence of the intelligent system. This paper discusses emphatically data mining techniques which could be used in CBR, and puts forward two algorithms of case base acquisition from historical data base automatically. The experimental result shows that this idea speeds up the implementation cycle of CBR intelligent system and strengthens the running efficiency of the system.
出处
《天津大学学报（自然科学与工程技术版）》
EI
CAS
CSCD
北大核心
2003年第1期23-27,共5页
Journal of Tianjin University：Science and Technology
基金
国家自然科学基金资助项目(60075015)
安徽省教育厅自然科学基金资助项目(2001kj002)
