基于MapReduce的移动社交网络下短文本内容相似性Top-k时空查询算法
Top-k space-time query algorithm for short text content similarity based on MapReduce in mobile social networks
投稿时间:2018-05-11  修订日期:2018-05-11
DOI:
中文关键词: 社交网络  海量短文本  滑动窗口  分布式查询
英文关键词: social network  massive short text  sliding window  distributed query
基金项目:国家科技支撑计划项目子课题(2015BAD29B01)
作者单位E-mail
雷建云 中南民族大学 计算机科学学院 leijianyun@mail.scuec.edu.cn 
彭媛 中南民族大学 计算机科学学院  
孙翀 中南民族大学 计算机科学学院  
帖军 中南民族大学 计算机科学学院  
摘要点击次数: 73
全文下载次数: 
中文摘要:
      随着移动社交网络快速发展,在海量带有时间属性和地理位置属性的短文本信息中查询到用户需求的信息具有重要意义。移动社交网络下短文本内容相似性Top-k查询算法忽略时间维度,并且需要满足用户在海量数据下快速响应的需求。针对以上问题,设计了基于MapReduce框架下的查询框架提高用户查询结果质量,该框架使用了基于滑动窗口下的多版本时空索引,保证了查询过程中融合了时间和空间属性,提高用户查询结果质量,实现高效的对海量数据的分布式查询目标。最后,通过真实数据集的实验证明提出的查询框架保证了查询的高效性。
英文摘要:
      With the rapid development of mobile social networks, it is of great significance to search for information required by users in massive short text information with time attributes and geographic attributes. The short text content similarity in mobile social network Top-k query algorithm ignores the time dimension, and needs to meet the user''s demand for rapid response under massive data. Aiming at the above problems, a query framework based on MapReduce framework is designed to improve the quality of user query results. The framework uses a multi-version spatiotemporal index based on the sliding window to ensure that time and space attributes are integrated in the query process and the quality of user query results is improved. To achieve an efficient distributed query target for massive data. Finally, experiments on real life dataset prove that the proposed query framework guarantees the efficiency of the query.
View Fulltext   查看/发表评论  下载PDF阅读器
关闭