基于遗传算法的全局优化检索策略研究  

Strategy research of global optimization based on genetic algorithm

作　　者：李海芳[1] 樊海亮[1] 金辉[1] 

机构地区：[1]太原理工大学计算机与软件学院,太原030024

出　　处：《计算机应用研究》2009年第5期1667-1669,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(60773004);山西省自然科学基金资助项目(2006011030)

摘　　要：案例检索是基于案例推理(CBR)系统中的关键技术,也是实现智能挖掘系统的关键环节。为了能够进一步提高案例检索效率与准确性,传统研究多是从案例属性和案例库的约减两方面入手,但是没有考虑案例属性权重。提出了一种基于遗传算法的全局优化案例检索模型,该模型利用遗传算法在搜索优化上的优势,对案例库、属性权重、K-NN中的K值进行全局同步优化。最后,通过实验验证了该模型在检索效率与准确性上优于传统模型。Case retrieval is the key technique in CBR,and it is also the most important step in realizing artificial mining system. To improve the efficiency and accuracy of case retrieval, most traditional researches focus on attribute reduction and filter casebase, but they do not take attribute weights and K of K-NN into account. This paper proposed a new global optimization research technique based on genetic algorithm. This technique, taking advantage of genetic algorithm, can optimize case-base, attribute weights, and K of K-NN algorithm synchronously. At last, experiments show that searching efficiency and accuracy under this technique are priorer than traditional technique in efficiency and accuracy.

关 键 词：K近邻搜索算法 遗传算法 案例检索 基于案例推理 全局优化 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]