基于CNM-BA的投标异常行为的离散化信息识别技术  

Discrete information recognition technology of abnormal bidding behavior based on CNM-BA

作　　者：宋永春 李志 曹雨 阚涛 宋树 SONG Yongchun;LI Zhi;CAO Yu;KAN Tao;SONG Shu(Anhui Wandian Tendering Co.,Ltd.,Hefei 230022 China;Anhui Jianzhu University,Hefei 230601,China;Anhui Jiyuan Software Co.,Ltd.,Hefei 230022,China)

机构地区：[1]安徽皖电招标有限公司,安徽合肥230022 [2]安徽建筑大学,安徽合肥230601 [3]安徽继远软件有限公司,安徽合肥230022

出　　处：《粘接》2024年第10期137-139,148,共4页Adhesion

基　　金：安徽省重大科技攻关资助项目(项目编号:202003A04080139)。

摘　　要：电力供应商围标串标主要依靠人工经验识别,准确率低、识别时间长和识别误差大。结合围标、串标中的离散化行为特征,设计基于社团检测-蝙蝠算法(CNM-BA)识别围标串标行为。融合改进蝙蝠算法和关联规则算法提取围标串标行为特征,离散化获取交易轨迹和行为特点,创新性地离散化改进蝙蝠算法;利用支持向量机算法对交易轨迹和特点信息进行识别分类,识别围标串标异常行为。实验结果表明,该方法的识别准确率高于88%,识别时间小于10 s,降低了识别误差,可以准确识别围标串标行为。The power supplier’s bid collusion mainly relies on manual experience identification,with low accura⁃cy,long identification time and large identification error.Combined with the discrete behavior characteristics of bid collusion and bid collusion,the design was based on the Finding Local Community Structure in Networks-Bat Algorithm(CNM-BA)to identify the collusion behavior.The improved bat algorithm and association rule algorithm were integrated to extract the behavior characteristics of the beacon collusion,the transaction track and behavior characteristics were discretized,and bat algorithm was creatively discretely improved.The support vector machine algorithm was used to identify and classify the trading track and characteristic information,and to identify the abnormal behavior of bid collusion.The experimental results showed that the recognition accuracy of this method was higher than 88%,and the recognition time was less than 10 s,which reduces the recognition error and can accurately identify the behavior of collusion.

关 键 词：CNM-BA离散化 围标串标 蝙蝠算法 关联规则 支持向量机算法 

分 类 号：TP391.97[自动化与计算机技术—计算机应用技术]