支持联机分析处理的推特用户兴趣维层次提取方法

发布时间：2018-05-05 08:39

本文选题：联机分析处理 + 推特　；参考：《电子与信息学报》2017年09期

【摘要】：从海量推特数据中探索用户兴趣的分布规律和相关性有利于实现精确的个性化推荐。联机分析处理(OnLine Analytical Processing,OLAP)提供了一种适合人们探究数据的直观形式。将OLAP技术应用于推特数据的关键是如何挖掘和构建推特用户的兴趣维层次。针对现有方法只能提取单一层次兴趣的不足,该文提出一种支持联机分析处理的推特用户兴趣维层次提取方法。该方法首先通过Rest API获取推特数据,然后通过改进的LDA(Latent Dirichlet Allocation)模型挖掘用户的兴趣和子兴趣,最后在此基础上构建兴趣维层次结构。实验评估了该方法的模型效果和可扩展性,并证实与LDA和h LDA相比可以更有效地提取出推特用户的兴趣维层次并应用于联机分析处理。
[Abstract]:Exploring the distribution and correlation of user interest from massive Twitter data is beneficial to the realization of accurate personalized recommendation. Online Analytical processing (OLAP) provides an intuitive form for people to explore data. The key to apply OLAP technology to Twitter data is how to mine and construct the interest dimension of Twitter users. In view of the disadvantage that the existing methods can only extract interest at a single level, this paper proposes a method for extracting the dimension of interest of Twitter users, which supports on-line analytical processing (OLAP). The method firstly acquires Twitter data through Rest API, and then uses the improved LDA(Latent Dirichlet allocation model to mine the interest and sub-interest of users. Finally, the hierarchy of interest dimension is constructed on the basis of this. The model effect and extensibility of this method are evaluated and compared with LDA and h LDA, it is proved that this method can extract the interest dimension of Twitter users more effectively and be applied to OLAP.
【作者单位】：杭州电子科技大学计算机学院;
【基金】：国家自然科学基金项目(61100043,61472112) 浙江省自然科学基金资助项目(LY12F02003) 浙江省科技计划重点资助项目(2017C01010,2016F50014)~~
【分类号】：TP391.1

【相似文献】