摘要
为提高相似案例选择的效率和准确性,将有向无环图支持向量机(DAGSVM)多类分类器应用到相似案例选择中,提出多类分类器有效分辨阈值的概念,在保证一定案例选择准确度的前提下,对自适应构造案例集进行相似案例选择,提高相似案例选择的效率.将该方法应用于光动力治疗(PDT)鲜红斑痣(PWS)案例推理专家系统,实验结果表明了该方法的有效性.
In order to improve the accuracy and efficiency of retrieving in case-based reasoning (CBR), an algorithm based on case retrieval is proposed, which uses directed acyclic graph support vector machines (DAGSVM) in retrieval. The effective differentiation threshold of DAGSVM is defined, and an adaptive subset of cases is built for each new case and the similar cases are selected in the subset. Using the approach, case retrieval is more efficient and accurate. The approach is applied in photodynamic therapy port wine stain CBR system, and the results show a dramatic increase in retrieving efficiency.
出处
《控制与决策》
EI
CSCD
北大核心
2007年第3期357-360,共4页
Control and Decision
关键词
基于案例推理
案例选择
支持向量机
有向无环图
Case-based reasoning
Case retrieval
Support vector machines
Directed acyclic graph