摘要
现有相似度量方法未考虑关键属性的否决作用与虑属性值裕度对案例推理的影响。为此,提出一种新的案例相似度量方法。定义属性值裕度和关键属性否决权,采用布尔函数、熵值理论和Composite策略等数学理论和分析工具,设计否决权和裕度属性的相似度量公式,给出先局部后全局的合并策略。实例结果表明,该方法在保证相似度量效能的前提下,能提高案例相似度量的分辨率。
Current similarity computation methods consider neither decisive function of key attributes,nor influence to case reasoning that is aroused by attribute value margin,a new case similarity computation method is proposed.It includes the innovations,such as,defining the attribute value margin and key attribute,designing new computation formulas and realizing method that computes whole case similarity after part similarity computing.Boolean function,entropy theory and composite combination strategy are used as math theory and analysis tools.And on precondition of having the same similarity computation efficiency,results prove that the new method can advance similarity computation resolving power ground.
出处
《计算机工程》
CAS
CSCD
2012年第12期261-264,共4页
Computer Engineering
基金
国家"973"计划基金资助项目(613900201)
空军工程大学导弹学院研究生创新基金资助项目(HX1112)
关键词
相似度量
属性值裕度
否决权
布尔函数
熵值理论
合并策略
similarity computation
attribute value margin
veto power
boolean function
entropy theory
combination strategy
