面向简化决策分析的概念格约简方法研究

项目名称： 面向简化决策分析的概念格约简方法研究

项目编号： No.61305057

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 李金海

作者单位： 昆明理工大学

项目金额： 23万元

中文摘要： 应用形式概念分析从复杂数据中获取知识进行决策分析是当前知识工程领域的一项重要任务，而决策分析的效率和效果是衡量这一应用在多大程度上取得成功的两个重要指标。本项目将围绕从复杂数据中获取知识进行决策分析的效率和效果如何提高展开研究。具体内容：采用“分而治之”策略提取决策形式背景的蕴含规则，且引入论域压缩技术设计概念格约简算法，从而提高决策形式背景在约简前后进行决策分析的效率；在多源形式背景中构造概念格以实现多源信息融合，研究面向简化决策分析的概念格约简以得到多源决策形式背景的决策分析方法；从决策分析的角度对现有的主要概念格约简方法进行比较研究，并对它们实施泛化和特化操作形成新的概念格约简方法，在此基础上给出一个完备格结构揭示这些约简方法之间的关联性，进而建立系统化的面向简化决策分析的概念格约简框架。本项目的研究将得到一些可行且有效的从复杂数据中获取知识进行决策分析的形式概念分析方法。

中文关键词： 形式概念分析；概念格；决策形式背景；决策分析；约简

英文摘要： In the field of knowledge engineering, knowledge acquisition from complex data for decision-making analysis has become one of the main tasks to formal concept analyis in which efficiency and effectiveness are two important indices for measuring how much this application can be successful. In order to improve these two indices of decision-making analysis based on the knowledge discovered from a complex data, we make a series of studies. Concretely, there are three aspects to be investigated: a divide-and-conquer strategy is used to extract implication rules from a decision formal context and the technique of compressing universe of discourse is employed to design a reduction algorithm of concept lattice, thereby improving the efficiency of decision-making analysis for decision formal contexts before and after the reduction, respectively; construct concept lattice of a multi-source formal context for multi-source information fusion, and study the issue of decision-making analysis oriented concept lattice reduction to obtain a decision-making analysis method of a multi-source decision formal context; from the perspective of decision-making analysis, a comparative study is made among the main existing concept lattice reduction methods that are generalized and specialized to obtain new concept lattice reduction meth

英文关键词： Formal concept analysis；Concept lattice；Decision formal context；Decision-making analysis；Reduction