首页 | 本学科首页   官方微博 | 高级检索  
     

类比转换原理及其实现
引用本文:李波,罗玉龙,赵沁平.类比转换原理及其实现[J].软件学报,1995,6(3):164-172.
作者姓名:李波  罗玉龙  赵沁平
作者单位:北京航空航天大学计算机系,北京,100083;北京航空航天大学计算机系,北京,100083;北京航空航天大学计算机系,北京,100083
基金项目:本研究受到国家自然科学基金和863计划的资助.
摘    要:类比转换完成将已知情况(称基)的知识引入到相似新情况(称靶),从而求解靶或学习到关于靶的新知识.本文的类比转换原理讨论了如何选择最佳映射,怎样在靶中创建对象和谓词,以及转换基中那些命题到靶.并基于该原理设计了类比转换的计算模型,实现了类比转换器ATE.实例分析表明ATE生成的类比结论既具创造性,又有较高可信度.

关 键 词:类比转换,类比推理,机器学习,类比学习
收稿时间:1992/6/18 0:00:00
修稿时间:1993/9/10 0:00:00

THE PRINCIPLES AND IMPLEMENTATION OF ANALOGICAL TRANSFER
Li Bo,Luo Yulong and Zhao Qinping.THE PRINCIPLES AND IMPLEMENTATION OF ANALOGICAL TRANSFER[J].Journal of Software,1995,6(3):164-172.
Authors:Li Bo  Luo Yulong and Zhao Qinping
Abstract:Analogical transfer process carries knowledge from a known situation(the base) over to a similar new situation(the target) so that one can solve the target or obtain new knowledge about the target.The principles of analogical transfer in this paper discuss how to select the best mapping,how to create objects and predicates in the target, and which proposition of the base to be transferred into the target. On the basis of the principles, the paper designs a computational model of analogical transfer and implements an analogical transfer engine called ATE.The case study illustrates analogical conclusions produced by ATE are with both creativity and higher certainty degree.
Keywords:Analogical transfer  analogical reasoning  machine learning  analogical learning  
本文献已被 CNKI 维普 等数据库收录!
点击此处可从《软件学报》浏览原始摘要信息
点击此处可从《软件学报》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号