网络用户行为大数据采集分析引擎设计  

Design of Big Data Collection and Analysis Engine for Network User Behavior

作　　者：赵露[1] 李庐[2] ZHAO Lu;LI Lu(Anhui Vocational College of Electronics&Information Technology,Bengbu,Anhui 233000;Anhui University of Finance and Economics,Bengbu,Anhui 233000)

机构地区：[1]安徽电子信息职业技术学院,安徽蚌埠233000 [2]安徽财经大学,安徽蚌埠233000

出　　处：《绵阳师范学院学报》2022年第11期83-87,共5页Journal of Mianyang Teachers＇ College

基　　金：安徽省高校优秀青年人才支持计划项目(gxyq2021282);安徽省质量工程教学研究项目(2020jyxm0133);安徽省质量工程重大教学研究项目(2020jyxm0121);安徽省质量工程重点教学研究项目(2020jyxm0123)。

摘　　要：由于网络用户行为数据的规模较大,难以实现对其完整采集,导致用户行为分析准确性较低,为此,提出网络用户行为大数据分析引擎设计.以具备灵活的软硬件全可编程能力和并行算法FACE-ZU-15EG作为引擎的数据采集构件,在板载FPGA协处理器的支持下实现对数据采集速度的适应性调节,确保采集结果的完整性,通过PageRank计算出高活跃度的用户,以其为中心,利用HITS算法计算关联用户聚类的行为特征.测试结果表明,设计方法可以实现对数据的有效采集,分析准确性在70%以上.Due to the large scale of network user behavior data, it is difficult to collect it completely, resulting in low accuracy of user behavior analysis. Therefore, the design of network user behavior big data analysis engine is proposed. Using FACE-ZU-15 EG with flexible software and hardware programmability and parallel algorithm as the data acquisition component of the engine, with the support of the onboard FPGA coprocessor, the data acquisition speed can be adaptively adjusted to ensure the integrity of the acquisition results. To calculate the high activity of users through PageRank, use the HITS algorithm to calculate the behavior characteristics of associated user clusters. The test results show that the design method can effectively collect data, and the analysis accuracy is above 70%.

关 键 词：网络大数据 用户行为 分析引擎 FPGA协处理器 HITS算法 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]