物联网差异数据库中的故障数据快速挖掘仿真  

Differences Between the Internet of Things in the Database Failure Data Rapid Excavation Simulation

作　　者：钱晓军[1] 范冬萍[1] 吉根林[1] 

机构地区：[1]南京师范大学计算机学院,江苏南京210097

出　　处：《计算机仿真》2016年第1期301-304,共4页Computer Simulation

基　　金：江苏省科技支撑计划(BE2013685);江苏省前瞻性联合研究计划项目(BY2014001-03)

摘　　要：对大型物联网环境下的差异数据库中的故障数据进行准确检测时,由于大型物联网环境下差异数据库受到环境状况和数据质量的影响,造成故障数据挖掘时数据特征描述不准确。传统的算法在故障数据挖掘过程中,由于受到不准确特征的影响,无法准确划分故障数据与非故障数据的邻域,导致挖掘准确率降低。提出采用离群点算法的大型物联网环境下的差异数据库中的故障数据挖掘方法。将差异数据库中的数据属性分为故障属性和非故障属性,用故障属性计算故障邻域,用非故障属性确定数据的故障离群度,对差异数据库中的所有的数据的故障离群程度进行计算,并进行降序排列,从而完成故障数据的挖掘。实验结果表明,利用改进算法进行差异数据库中的故障数据挖掘,能够有效提高挖掘的准确率,降低挖掘耗时。For large differences under the environment of Intemet of things in the database failure data accurately digging, differences due to the large Internet environment database affeeted by environmental conditions and the data quality, cause the fault data mining data character description is not accurate. The traditional algorithm in fault in the process of data mining, due to the effect of the inaccurate features cannot be accurately classified fault data and fault data of neighborhood, results in the decrease of mining accuracy. Outliers based on the algorithm of the large differ- ences in under the environment of Internet of things in the database failure data mining method. Difference attribute data in the database can be divided into fault and the fault attribute, use fault attribute evaluation neighborhood, with a fault properties determine the fault degrees from the group of data, the differences between all of the data in the da- tabase of fault degree is calculated from the group, and in descending order, to complete the failure data mining. The experimental results show that the improved algorithm difference fault data mining in the database, can effectively im- prove the accuracy in mining, reduce mining time-consuming.

关 键 词：大型物联网 差异数据库 故障数据 挖掘 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]