摘要
知识表示与类比推理的一体化研究是将类比推理从理论应用于实践的关键。类比推理所依倨的知识的主要特征是知识的多层次性和不完全性，将神经网络与事例推理引入类比推理可以自然地体现这种特征并高效地实现类比推理。利用模糊集与模糊逻辑可以对类比对象间的有条件性相似关系做合理的数学描述，准确确定类比的条件规则。通过上述工作我们使类比推理从传统的抽象逻辑推理转向基于知识的推理，为其进一步的实用化研究提供新的理论和方法。
The research on the integration of knowledge representatiom and analogy reasoning is the very crux that turns analogy reasoning from theory into practice. The main character of the knowledge on which analogy reasoning depends is the multilayeral and incomplete data.Neural network and case-based reasoning can naturally represent this character and efficently make analogy reasoning. Fuzzy set and fuzzy logic can rationally make mathematical description of the conditional similarity among the analogical objects and correctly determine the rule of the analogy. By these, we turn the traditional analogy from abstract logic reasoning to knowledge-based reasoning and support new theories and methods for the practical research of analogy reasoning.
基金
国家自然科学基金
关键词
类比推理
知识库
神经网络
事例推理
模糊数学
analogy reasoning
knowledge base
neural network
case-based reasoning
fuzzy logic and fuzzy set
