基于地理近邻关系的微博系统朋友推荐

发布时间：2018-06-23 13:54

  本文选题：微博系统 + 朋友推荐　； 参考：《计算机工程与应用》2017年13期

【摘要】：近年来,微博的蓬勃发展吸引了大量网络用户,用户所发海量微博呈现的大数据环境成为理解用户行为的重要资源。目前,大量在线朋友推荐研究通过对微博内容分析推断用户的兴趣和喜好以进行朋友推荐,但大多数已有研究忽略了用户位置和兴趣之间的潜在关系。事实上,多数情况下用户真正感兴趣的还是他周围的人。为此,提出了基于地理近邻关系的朋友推荐方法,通过把所处位置周围兴趣爱好相似的微博用户彼此推荐,为用户提供了与周围可能感兴趣的人联系的独特渠道。仿真分析证明,与传统朋友推荐方法相比,基于地理近邻的朋友推荐具有较高的推荐性能。
[Abstract]:In recent years, the vigorous development of Weibo has attracted a large number of network users. The big data environment presented by massive Weibo issued by users has become an important resource to understand user behavior. At present, a large number of online friend recommendation studies infer the interests and preferences of users by analyzing the content of Weibo, but most of the previous studies have neglected the potential relationship between user location and interest. In fact, in most cases the user is really interested in the people around him. In this paper, a method of recommending friends based on geographical nearest neighbor relationship is proposed. By recommending Weibo users with similar interests and hobbies around their location, it provides a unique channel for users to communicate with people who may be interested in their surroundings. The simulation results show that compared with the traditional friend recommendation method, the friend recommendation based on geographical nearest neighbor has higher recommendation performance.
【作者单位】： 天津师范大学计算机与信息工程学院;电子科技大学计算机与工程学院;
【基金】：国家自然科学基金(No.61103227,No.61272526,No.61472068,No.61572113) 中国博士后基金(No.2014M550466,No.2014M562308,No.2014M562310)
【分类号】：TP391.3;TP393.092
，

本文编号：2057335