项目名称: 基于公理模糊集理论的模糊机器学习
项目编号: No.60875032
项目类型: 面上项目
立项/批准年度: 2009
项目学科: 数理科学和化学
项目作者: 刘晓东
作者单位: 大连理工大学
项目金额: 28万元
中文摘要: 基于人类感知和认知方法,借鉴经典机器学习理论和技术,在公理模糊集理论框架内,应用测度论、格论、拓扑学、组合学和现代数据统计分析方法建立一套系统且较完善的模糊机器学习理论和方法。其算法不仅可直接应用许多深刻抽象的数学理论和计算机进行深入地研究、分析和处理,而且可用人类自然语言解释和思维逻辑理解与分析。这一方面克服目前模糊机器学习面临"黑盒子"问题,另方面通过模仿人类感知和认知,建立能够用自然语言、思维逻辑和经验直接解释的基于数据的知识结构。其研究成果在数据可解释性建模、基于数据驱动的控制、知识发现与表示等方面具有重要的理论和应用价值。最后相应的算法将应用机器人进行验证并提高其避让障碍物和路径规划的智能水平和速度。
中文关键词: 机器学习;模糊概念表示;模糊逻辑运算;公理模糊集理论
英文摘要: Based on human perception and recognition methods, using classical machine learning theory and techniques, under the framework of Axiomatic Fuzzy Sets theory, apply measure theory, lattice theory, topology, combinatorics and modern statistic analysis to establish a systems of fuzzy machine learning theory and method. The algorithms not only can be analyzed by some abstract mathematic theories and investigated by computers, but also can be comprehended by natural language and analyzed by fuzzy logic. Thus, this approach will achieve two goals: one is that it can overcome the "black box" problems in fuzzy machine learning; another one is that it can establish a data based knowledge structure with natural language, thinking logic and experiment interpretation through imitating human perception and recognition. The results will be found important applications and theory values in data interpretable modeling, data driven control and knowledge discovery and representation, etc. Finally, some algorithms will be verified on the applications to improve the intelligent level of the robots route designs.
英文关键词: Machine Learning; Fuzzy Concept Representation; Fuzzy Logical Operations; Axiomatic Fuzzy Sets Theory