并行程序中多路径数据相似性检测方法研究

Research on similarity detection of multipath data in parallel programs

针对传统多路径数据相似性检测方法查准率与检测速度低的情况,对并行程序中多路径数据相似性检测方法进行设计。按照数据属性对多路径数据集划分,并对数据预处理,根据预定义识别相似规则,判断多路径数据库中的数据是否存在相似重复记录,以此完成对多路径数据相似性检测。实验结果表明,在实验数据55.2万时,传统方法的查准率比此次设计方法查准率低6.2%,在实验数据385.2万时,传统方法比此次方法查准率低17.9%。并且在相同查全率条件下,在实验数据55.2万时,传统方法时间比此次设计的方法时间多10.1 min,在实验数据385.2万时,传统方法检测时间比此次设计方法检测时间多21.2 min。说明,此次设计的检测方法查准率比传统方法查准率高,并且在相同查全率的条件下,传统方法的检测时间比本文时间耗费多,因此说明该方法查准率与检测速度高。

Aiming at the low precision and detection speed of traditional multi-path data similarity detection method,the multi-path data similarity detection method in parallel program is designed.According to the data attributes,the multipath data sets are divided,and the data are preprocessed.According to the predefined recognition similarity rules,the existence of similar duplicate records in the multipath database is judged,so as to complete the similarity detection of multipath data.The experimental results show that the accuracy of the traditional method is 6.2%lower than that of the design method when the experimental data are 552000,and 17.9%lower than that of the current method when the experimental data are 385000.Under the same recall rate,when the experimental data is 552000,the time of traditional method is 10.1 min longer than that of this design.When the experimental data is 385000,the time of traditional method is 21.2 min longer than that of this design method.It shows that the precision rate of the proposed method is higher than that of the traditional method,and under the same recall rate,the detection time of the traditional method is more time-consuming than that of this paper,so it shows that the precision rate and detection speed of this method are higher.