案例系统特征权值优化算法的仿真研究  

Simulation research on case system feature weights optimization algorithm

作　　者：王观玉[1] 郭勇[1] 

机构地区：[1]黔南民族师范学院计算机科学系,贵州都匀558001

出　　处：《计算机工程与应用》2013年第1期261-264,共4页Computer Engineering and Applications

摘　　要：研究案例系统特征权值优化问题,传统特征权值确定方法过分依赖主观判断和经验,而单一遗传算法或禁忌算法存在各自的不足,因此案例分类精度低。为了提高案例分类精度,提出一种遗传算法和禁忌算法相融合的案例系统特征权值优化方法。利用遗传算法全局搜索能力、并行性和禁忌算法局部搜索和记忆能力,有效地解决了案例系统特征权值优化问题。仿真结果表明,混合方法利用了遗传算法和禁忌算法的优点,很好地优化了案例系统特征权值,从而加快案例系统检索速度,提高了案例分类精度。Study about case system feature weights optimization problem, the traditional method feature weights＇ determininatiion has the over-reliance on subjective judgment and experience, and a single Genetic Algorithm （GA） or Tabu Algorithm （TA） has their shortcomings, case classification accuracy is low. In order to improve the case classification accuracy, this paper proposes a case system feature weights optimization method based on GA and TA. Using GA global search capability, parallelism and TA local search and memory capacity, it effectively solves the case system feature weights optimization problems. Simulation re-sults show that the hybrid method using the GA and TA merit, can optimize case system feature weights, speed up case retrieval speed and improve the system of case classification accuracy.

关 键 词：案例系统 特征权值 遗传算法 禁忌算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]