全概念类比中元知识结构表示的研究

项目名称： 全概念类比中元知识结构表示的研究

项目编号： No.61272545

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 沈夏炯

作者单位： 河南大学

项目金额： 80万元

中文摘要： 目前类比学习模型依赖于人工建立的概念及其联系，不能利用数据中隐含的未标识概念，在知识的获取完整性和表示完备性两方面均不利于海量数据的自动处理。本项目基于形式背景核的概念格同构方法，拟研究一种元知识结构表示方法和一种从数据中自动计算全部概念并加以利用的分布式网络计算模型。首先将各种形态的形式背景集合抽象成有限的n 阶形式背景核，以其概念格簇表示元知识，形成知识内核；其次利用形式背景同构判定算法，寻找从元概念格到靶空间的映射，通过映射完成元知识的迁移，形成新的知识，藉此实现自动类比学习；最后构建计算模型，由元知识库和一组实现类比的机制构成，元知识库由n 阶的形式背景核及其概念格组成，类比机制包括形式背景的同构判定、概念格的同构等一组算法。本项目重点提高类比学习模型的完备性和自适应性，研究结果为知识发现的数据自动处理提供算法模型和理论依据, 对促进知识学习和发现应用具有重要的理论意义和实用价值。

中文关键词： 形式背景核；概念；元知识；类比；同构

英文摘要： The analogical learning model relies on artificial building concept and contact, cannot use unlabeled data implied in the acquisition of knowledge concept, integrity and completeness of said two were detrimental to the automatic processing of mass data. The project is based on context kernel concept lattice isomorphism method, is proposed to research a yuan of structure of knowledge representation method and a data from the automatic calculation of all concepts and make use of the distributed network computing model. The various forms of abstract background set into a finite n th-order contexts kernel, with its concept lattice cluster represents knowledge, form the kernel of knowledge; then use context isomorphism detection algorithm, looking from the element concept lattice to the target space mapping, through the mapping to complete the element of knowledge transfer, the formation of new knowledge, to realization of automatic learning by analogy; finally constructs the calculation model, the element knowledge base and a set of analogy mechanism composition, element knowledge base by n order form background nucleus and its concept lattice, analogy mechanism includes formal context, concept lattice isomorphism isomorphism and a group of algorithms. The emphasis of this project to improve the completeness of anal

英文关键词： Context kernel；Concept；Meta-knowledge；Analogical；Isomorphism