基于属性约简的GEP洪灾灾情预测算法  

Flood Forecast Algorithm for GEP Based on Attribution Reduction

作　　者：杨乐婵[1] 邓岳川[1] 邓松[2] 谢昀[2] 

机构地区：[1]滁州学院国土信息工程系,安徽滁州239000 [2]南京邮电大学计算机学院,江苏南京210003

出　　处：《计算机技术与发展》2009年第9期87-89,92,共4页Computer Technology and Development

基　　金：安徽省青年教师科研计划项目(2008jq1115)

摘　　要：洪灾灾情预测是预防和了解洪灾的重要手段。为了解决传统的洪灾灾情预测算法的不足,提出基于折半查找的最优属性约简算法(OAR-BSA),最大限度地降低影响洪灾灾情数据的因素,并在此基础上,结合基因表达式编程的概念,给出了基于属性约简的GEP洪灾灾情预测算法(FFGEP-AR)。仿真实验表明,OAR-BSA算法求解最优约简比传统的算法要快,FFGEP-AR算法比传统预测算法的平均耗时要小,预测的准确率要高。Flood forecast is an important method of preventing and understanding flood. To solve the problem of traditional flood forecast algorithms, presents optimum attribution reduction on binary search algorithm （OAR- BSA） in order to reduce factors on affecting flood disaster data. Based on OAR-BSA, flood forecast for GEP on attribution reduction （FFGEP- AR） is proposed. By extensive experiments, it is showed that OAR- BSA is faster than traditional algorithms on solving optimum attribution reduction, average consumptive time of FFGEP-AR is less than traditional algorithms, and predictive accuracy is higher.

关 键 词：属性约简 基因表达式编程 灾情预测 函数挖掘 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]