基于位置社交网络的个性化推荐方法的研究
发布时间：2018-08-03 12:07
【摘要】：随着互联网技术的进一步发展,社交网络也在人们的生活中越来越流行起来。人们通过移动社交网络发现有用的生活信息,分享生活经验,与好友互动。并且伴随着2010年前后安卓,苹果等智能移动设备的应用普及。在原有社交网络的基础上结合了用户地理位置的相关信息,形成了一个新的社交网络概念:基于位置的社交网络(Location-based Social Network,LBSN)。基于位置的社交网络不仅关注用户线上发表的信息和用户线上的好友关系,还保留了用户线下的活动数据与活动模式,了解用户的日常行为。由于LBSN包含了大量的多种类型的数据。我们可以对这些数据进行挖掘,发现有意义的信息。目前,对于利用LBSN数据对用户进行个性化推荐的研究热度非常高。利用LBSN数据对用户推荐主要集中在以下三个领域:(1)兴趣点推荐:这一推荐是针对用户的,包括两种推荐,一种是目的地点,另一种是多个地点结合的路线推荐;(2)商业选址推荐:这种推荐主要针对商户;(3)好友推荐:对用户进行线上好友推荐。尽管学者们对LBSN推荐系统的研究有了很大的突破,但目前还存在着以下一些问题:(1)兴趣点与用户众多,计算量大,数据稀疏;(2)用户关系表示方法过于简单,包含内容不多;(3)新加入用户的冷启动问题(4)没有结合时空上下文信息等。针对以上问题,本文提出了一些新的想法与解决方案(1).通过划分城市单元格,缩减问题分析规模,解决部分数据稀疏问题;(2)将好友关系的表示丰富起来,更加符合现实世界。在普遍的LBSN数据中,好友关系只用01(有无)表示,显然没有考虑到好友关系的亲密度,本文会结合线上线下数据,对好友关系进行量化,从而对推荐系统起到更有意义的作用。(3)针对上下文信息的问题,本文提出结合时空信息,提出结合用户时空特征的推荐,分析用户与兴趣点的时空特征以及当前时空信息来做更好的推荐。另外,本文会结合线上信息和线下信息,考虑到实际区域的一些地理属性,提出一种新的商业选址问题的想法和一种基于用户与兴趣点之间随机游走的好友推荐模型。在试验中,我们选取国外比较著名的LBSN数据集Foursquare进行分析与实验,通过准确率,召回率等一些评判标准比较文中的方法和之前的研究成果。对文章的算法进行一系列测试,能实现对用户和POI供应商的有效推荐。
[Abstract]:With the further development of Internet technology, social networks are becoming more and more popular in people's lives. People use mobile social networks to find useful life information, share life experiences, and interact with friends. And with Android, Apple and other smart mobile devices around 2010, the popularity of applications. Based on the existing social network, a new concept of social network, Location-based Social Network (LBSN), is formed by combining the relevant information of the user's geographical location. The location-based social network not only pays attention to the information published on the user line and the friends relationship on the user line, but also preserves the activity data and the activity pattern below the user line to understand the daily behavior of the user. Because LBSN contains a large number of types of data. We can mine the data and find meaningful information. At present, the research on personalized recommendation based on LBSN data is very hot. Using LBSN data to recommend users is mainly focused on the following three areas: (1) Point of interest recommendation: this recommendation is for users, including two types of recommendations, one is the destination point, The other is the route recommendation of multiple locations; (2) business location recommendation: this recommendation is mainly for merchants; (3) friend recommendation: online friend recommendation to users. Although scholars have made a great breakthrough in the research of LBSN recommendation system, there are still some problems as follows: (1) the number of interest points and users is large, the amount of computation is large, the data is sparse; (2) the expression method of user relationship is too simple. (3) the cold start problem of the new user (4) not combining the temporal and spatial context information. In view of the above problems, this paper puts forward some new ideas and solutions (1). By dividing urban cells, reducing the size of problem analysis, and solving the problem of partial data sparse; (2) enriching the representation of friends, more in line with the real world. In the general LBSN data, the friend relationship is only 01 (whether there is or not), and obviously does not take into account the intimate density of the friend relationship, this paper will combine the on-line and offline data to quantify the friend relationship. Therefore, it plays a more important role in recommendation system. (3) aiming at the problem of context information, this paper proposes to combine space-time information with user space-time features. Analyze the temporal and spatial characteristics of users and interest points and the current spatiotemporal information to make better recommendations. In addition, this paper proposes a new business location problem and a friend recommendation model based on random walk between users and points of interest. In the experiment, we choose the famous foreign LBSN data set Foursquare for analysis and experiment, and compare the methods and previous research results with some evaluation criteria such as accuracy, recall rate and so on. A series of tests on the algorithm of this paper can realize the effective recommendation to users and POI suppliers.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.3
[Abstract]:With the further development of Internet technology, social networks are becoming more and more popular in people's lives. People use mobile social networks to find useful life information, share life experiences, and interact with friends. And with Android, Apple and other smart mobile devices around 2010, the popularity of applications. Based on the existing social network, a new concept of social network, Location-based Social Network (LBSN), is formed by combining the relevant information of the user's geographical location. The location-based social network not only pays attention to the information published on the user line and the friends relationship on the user line, but also preserves the activity data and the activity pattern below the user line to understand the daily behavior of the user. Because LBSN contains a large number of types of data. We can mine the data and find meaningful information. At present, the research on personalized recommendation based on LBSN data is very hot. Using LBSN data to recommend users is mainly focused on the following three areas: (1) Point of interest recommendation: this recommendation is for users, including two types of recommendations, one is the destination point, The other is the route recommendation of multiple locations; (2) business location recommendation: this recommendation is mainly for merchants; (3) friend recommendation: online friend recommendation to users. Although scholars have made a great breakthrough in the research of LBSN recommendation system, there are still some problems as follows: (1) the number of interest points and users is large, the amount of computation is large, the data is sparse; (2) the expression method of user relationship is too simple. (3) the cold start problem of the new user (4) not combining the temporal and spatial context information. In view of the above problems, this paper puts forward some new ideas and solutions (1). By dividing urban cells, reducing the size of problem analysis, and solving the problem of partial data sparse; (2) enriching the representation of friends, more in line with the real world. In the general LBSN data, the friend relationship is only 01 (whether there is or not), and obviously does not take into account the intimate density of the friend relationship, this paper will combine the on-line and offline data to quantify the friend relationship. Therefore, it plays a more important role in recommendation system. (3) aiming at the problem of context information, this paper proposes to combine space-time information with user space-time features. Analyze the temporal and spatial characteristics of users and interest points and the current spatiotemporal information to make better recommendations. In addition, this paper proposes a new business location problem and a friend recommendation model based on random walk between users and points of interest. In the experiment, we choose the famous foreign LBSN data set Foursquare for analysis and experiment, and compare the methods and previous research results with some evaluation criteria such as accuracy, recall rate and so on. A series of tests on the algorithm of this paper can realize the effective recommendation to users and POI suppliers.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.3
【相似文献】
相关期刊论文 前10条
1 Bruce Antelman;李雯;;社交网络[J];高校图书馆工作;2008年01期
2 ;基于位置的手机社交网络“贝多”正式发布[J];中国新通信;2008年06期
3 曹增辉;;社交网络更偏向于用户工具[J];信息网络;2009年11期
4 ;美国:印刷企业青睐社交网络营销新方式[J];中国包装工业;2010年Z1期
5 李智惠;柳承烨;;韩国移动社交网络服务的类型分析与促进方案[J];现代传播(中国传媒大学学报);2010年08期
6 贾富;;改变一切的社交网络[J];互联网天地;2011年04期
7 谭拯;;社交网络:连接与发现[J];广东通信技术;2011年07期
8 陈一舟;;社交网络的发展趋势[J];传媒;2011年12期
9 殷乐;;全球社交网络新态势及文化影响[J];新闻与写作;2012年01期
10 许丽;;社交网络:孤独年代的集体狂欢[J];上海信息化;2012年09期
相关会议论文 前10条
1 赵云龙;李艳兵;;社交网络用户的人格预测与关系强度研究[A];第七届（2012）中国管理学年会商务智能分会场论文集（选编）[C];2012年
2 宫广宇;李开军;;对社交网络中信息传播的分析和思考——以人人网为例[A];首届华中地区新闻与传播学科研究生学术论坛获奖论文[C];2010年
3 杨子鹏;乔丽娟;王梦思;杨雪迎;孟子冰;张禹;;社交网络与大学生焦虑缓解[A];心理学与创新能力提升——第十六届全国心理学学术会议论文集[C];2013年
4 毕雪梅;;体育虚拟社区中的体育社交网络解析[A];第九届全国体育科学大会论文摘要汇编(4)[C];2011年
5 杜p，
本文编号：2161698
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2161698.html
