基于并行式频繁项集的党政收费平台

Party and government charge management platform based on parallel frequent itemset

为提高党政收费时效性与信息化管理水平,基于并行式频繁项集挖掘算法开发高效率、智能化的党政收费管理平台。基于云计算技术构建党政收费管理平台的总体架构,提供云缴费、党建教育学习、党建宣传等信息化功能。在Spark分布式计算框架上构建Spark集群,构造党政收费频繁项集挖掘矩阵,根据矩阵行列间运算获得频繁k项集支持度,利用“主-从”节点模式实现并行式频繁项集挖掘,获得党政收费管理信息分类结果。测试结果显示,该平台各功能最大平均响应时长仅为1.51 s,挖掘党政收费信息频繁项集的时间开销短、推荐非空率高,呈现了良好的频繁项集挖掘效率与质量。该平台助力优化党政费用交纳工作模式,为党员管理的信息化、智能化提供支持。

In order to improve the timeliness of party and government charges and the level of information management,an efficient and intelligent party and government charge management platform is developed based on the parallel frequent item set mining algorithm.The overall structure of the party and government fee management platform is constructed based on cloud computing technology to provide information functions such as cloud payment,party building education and learning,and party building publicity.The Spark cluster is constructed on the Spark distributed computing framework,and the mining matrix of frequent item sets of party and government charges is constructed.The support degree of frequent k item sets is obtained according to the calculation between the rows and rows of the matrix.The parallel mining of frequent item sets is achieved by using the“master-slave”node mode,and the classification results of party and government charges management information are obtained.The test results show that the maximum average response time of each function of the platform is only 1.51 s,and the time cost of mining frequent item sets of party and government fee information is short,and the recommendation non-empty rate is high,showing a good mining efficiency and quality of frequent item sets.The platform helps optimize the working mode of party and government expenses payment,and provides support for information-based and intelligent management of party members.