基于协同过滤技术的微博好友推荐方法的研究与实现

发布时间：2018-05-16 07:34

本文选题：社交网络 + 协同过滤　；参考：《中国海洋大学》2014年硕士论文

【摘要】：社交网络的迅速发展，给我们的生活带来了各项便利，使得越来越多的用户加入到其中来。社交网络软件也顺应潮流，源源不断地被开发出来。在社交网络中人们可以建立好友关系，通过好友关系查看好友的动态、日志等，也可以分享信息，实现即时通讯。社交网络以交友为核心，逐渐渗透到各个领域中，成为人类社会交流的一个工具。人们也正享受着这种“足不出户，能知天下事”的情景。现在的智能手机更是把社交网络展现的淋淋尽致，由于手机的普遍性和无线网络的应用，使得几乎每个人手里都攥着这个社交网络的载体。但是，面对如此多的用户，如何准确、高效找到与自己志趣相投的用户成了亟待解决的问题。好友推荐应运而生，解决了这一难题。本文在某政府内网建立的微博应用中设计实现了一种新的好友推荐方法，该方法整合政府所有人员的基本信息，政府人员发布的微博数目，以及应用中用户对推荐的反馈情况，实时的为用户推荐好友。既解决了用户繁多不知该添加谁的问题，又避免了新用户加入的尴尬情况，扩大了用户的交友圈。本文的主要研究内容如下：首先，介绍了推荐系统的概念、流程以及相关的理论，列举了几种常见的推荐技术，对不同推荐技术的优缺点也做了分析。其次，详细介绍了本论文使用的协同过滤技术，包括其核心思想、分类以及在不同分类下推荐算法的计算方法。并在此基础上，设计了一种新的好友推荐方法，对此方法的功能和实现流程做了详细设计。此方法的应用解决了以往的推荐算法对新用户，用户反馈等问题的疏忽，为政府内网微博的用户提供了准确的好友推荐功能。最后，将基于协同过滤技术的好友推荐方法用在已实现的政府内网微博中。该系统的客户端建立在当前最流行的Android系统上，实现了微博应用的相关功能；由服务器为客户端提供接口，主要实现数据存储和好友推荐两大功能模块。本文所实现的微博应用，使公务员可以随时随地的进行网上交流学习互动，脱离办公室电脑的束缚。系统除了实现现有的微博功能外，，还加入了好友推荐功能。该功能不仅仅依靠用户的个人信息推荐，还依据用户的微博使用以及用户对推荐的反馈情况。这不仅为公务员的学习交流提供更大的便利，同时也扩充了用户的使用量，更对用户扩大社交圈和学习圈提供更好的服务。
[Abstract]:With the rapid development of social network, more and more users join our daily life. Social networking software has also been developed in a steady stream. In the social network, people can establish a good friend relationship, check the friends' dynamic, log and so on through the good friend relationship, also can share the information, realize instant communication. Social network, which is centered on making friends, has gradually penetrated into various fields and become a tool for human social communication. People are also enjoying the sight of staying in the house and knowing the world. Today's smartphones are all about social networking. Because of the popularity of mobile phones and the use of wireless networks, almost everyone has the carrier of the social network in their hands. However, in the face of so many users, how to find users with similar interests accurately and efficiently becomes an urgent problem. Good friend recommendation arises at the historic moment, solved this difficult problem. In this paper, a new friend recommendation method is designed and implemented in the Weibo application set up by a government intranet. This method integrates the basic information of all government personnel, the number of Weibo released by government personnel, and the feedback from users on the recommendation in the application. Recommend friends for users in real time. It not only solves the problem of many users do not know who to add, but also avoids the awkward situation of new users to join, and expands the circle of users to make friends. The main contents of this paper are as follows: Firstly, this paper introduces the concept, flow chart and related theories of recommendation system, enumerates several common recommendation technologies, and analyzes the advantages and disadvantages of different recommendation technologies. Secondly, the cooperative filtering technology used in this paper is introduced in detail, including its core idea, classification and the calculation method of recommendation algorithm under different classification. On this basis, a new friend recommendation method is designed, and the function and implementation flow of this method are designed in detail. The application of this method solves the neglect of the previous recommendation algorithms to the new users, user feedback and so on, and provides the accurate friend recommendation function for the Weibo users of the government intranet. Finally, the best friend recommendation method based on collaborative filtering technology is used in the implemented Weibo. The client of the system is based on the most popular Android system, which realizes the related functions of Weibo application, and the server provides the interface for the client, mainly realizes the two function modules of data storage and friend recommendation. The application of Weibo in this paper enables civil servants to communicate and learn online anytime and anywhere, and break away from the shackles of office computers. The system not only realizes the existing Weibo function, but also adds the good friend recommendation function. This function depends not only on the user's personal information recommendation, but also on the user's Weibo usage and the user's feedback on the recommendation. This not only provides greater convenience for the learning and communication of civil servants, but also expands the usage of users, and provides better services for users to expand their social circle and study circle.
【学位授予单位】：中国海洋大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.092

【参考文献】