| 摘要: |
| 本文给出了一种类比匹配原理及其计算模型，并在此基础上实现了类比匹配器ＡＭＥ．其中类比匹配原理包括一致性、谓词相似和支持性原则；计算模型采用以命题为中心，以命题映射为基本单位，从而ＡＭＥ的复杂性低于同类系统．另外，证明了本文的实现算法具完备性和正确性． |
| 关键词: 类比匹配，类比推理，自动推理，机器学习 |
| DOI： |
| 分类号: |
| 基金项目:本文研究得到863计划和国家自然科学基金资助. |
|
| AN THEORY OF ANALOGICAL MATCH AND ITS IMPLEMENTATION |
|
Li Bo,Luo Yulong,Zhao Qinping
|
| Abstract: |
| This paper presents a theory and a computational model of analogical match, and an implementation of this model, named AME. The theory includes consistency, predicate similarity and support principles. Because the computational model constructs mapping in proposition-oriented match, and takes proposition mapping as primitive units, the complexity of AME is lower than the same kind of system. The paper also proves the implementation algorithms to be complete and reasonable. |
| Key words: Analogical match, analogical reasoning, automatic reasoning, machine learning. |
