基于大数据网络用户兴趣个性化推荐模型分析  

Analysis of personalized recommendation model of network user interest based on big data

作　　者：王蓉 李小青 刘军兰[1] 严晓梅[2] 陈瑜 WANG Rong;LI Xiao qing;LIU Jun lan;YAN Xiao-mei;CHEN Yu(Xi’an International University College of Engineering,Xi’an 710077,China;Airforce Engineering University Information and Navigation College,Xi’an 710077,China;Unit 95866 of People’s Liberation Army,Baoding 071000,China)

机构地区：[1]西安外事学院工学院,陕西西安710077 [2]空军工程大学信息与导航学院,陕西西安710077 [3]中国人民解放军95866部队,河北保定071000

出　　处：《电子设计工程》2019年第21期5-8,13,共5页Electronic Design Engineering

基　　金：国家自然科学基金青年科学基金项目(71503260)

摘　　要：针对传统分析方法受噪声和人为因素影响而造成分析结果较差的问题,我们提出了一种基于大数据的社交网络用户兴趣个性化推荐模型。在矢量空间模型的基础上,分析了用户兴趣推荐模型结构及其与周围模型的交互关系,划分了服务器网络部署模块,设计了运行模型网络结构。通过MapReduce模型将任务分布到分布式计算机集群中,用以构建用户感兴趣的个性化推荐模型。利用大数据双层关联规则数据挖掘技术获取用户感兴趣的网络数据,利用推荐结果确定用户对推荐内容的兴趣程度。实验对比结果表明,用此分析方法的分析效果可高达98%,对大规模社交网络用户的个性化推荐具有良好的可扩展性。Aiming at the problem that traditional analysis methods were affected by noise and human factors,resulting in poor analysis results,a method based on big data for social network user interest personalized recommendation model was proposed.Based on the vector space model,the user interest recommendation model structure and the interaction relationship with the surrounding model were analyzed,the server network deployment module was divided,and the operation model network structure was designed.The task was distributed to the distributed computer cluster through the MapReduce model to construct a personalized recommendation model for user interest.The big data double-layer association rule data mining technology was used to obtain the network data of the user interest,and the recommendation result was used to determine the degree of interest of the user to the recommended content.The experimental comparison results showed that the highest analysis effect reached 98%by using this analysis method,and it had good scalability for personalized recommendation of large-scale social network users.

关 键 词：大数据 社交网络 用户兴趣 个性化 推荐 模型 

分 类 号：TN915[电子电信—通信与信息系统]