摘要近年来,微博作为一种新兴的社交网络平台走入大众视野,并且迅速以其新颖便捷的信 息传播模式广受各类用户青睐。随着用户数量的大量增长，信息的总量也开始急剧增长, 如何从大量的微博信息中获取用户感兴趣的内容并推荐给用户,已经成为一个急需解决的 问题。用户建模分析用户兴趣和偏好的成功与否决定着推荐结果质量的好坏，只有对用户 的兴趣及偏好有了很好的理解，才能向用户推荐其满意的结果，而用户画像能快速找到精 准用户群体以及用户需求等更为广泛的反馈信息。国内外的研究主要是根据用户的浏览记 录、查询记录或检索词等来建立用户模型，而在用户兴趣获取、模型表示及评价等方面还 有不少缺陷。为了得到准确的微博用户兴趣模型，本文将基于微博内容和用户关系建立用 户兴趣模型。本文根据与中心用户共同关注的多少对他们进行重合度分析，取出最高的前 50 人，并对他们的标签进行词频统计，得出最多的 20 个标签，然后与用户自己的标签进 行相似度分析，建立用户模型。并设计实验来确定该微博用户兴趣模型的准确性，实验结 果表明，该模型能更加有效和准确地发现微博用户的兴趣。 79505

毕业论文关键词 微博  用户关系  微博内容  用户建模   个性化推荐  用户画像构建

Title User Modeling On Social Networks  Using User Relationship and Weibo Content for User Modeling       

Abstract Currently, Weibo came into the public view as a new social networking platform, and was widely popular with all types of users with its novel modes of information communication rapidly。 As the increase of the number of users, the amount of information has begun to rise sharply, how to obtain information that users interested in from a large number of weibo content and recommend to the user, has become an urgent problem。 The success of user modeling analysis of user's interests and preferences determine the quality of the recommended, only by understanding the user's interests and preferences well , can we recommended the satisfactory result to the user , User profilling can find the accurate user groups and user needs a broader feedback information quickly。 Research at home and abroad is mainly based on users' browsing record, query or retrieval words to build user model, but there are many shortcomings in the user interest acquisition, model representation and evaluation etc。 In order to get accurate user interest model, this paper will be based on the weibo content and user relationship to build user interest model。 In this paper, based on the amount and the users and coincidence degree analysis to them, take out the top 50 people, and make word frequency statistics for their labels ,take the top 20 tags, and then make similarity analysis with uaer's own labels to build the user model。 And design experiments to determine the accuracy of the model, the experimental results show that the model can found users' interest effectively and accurately。 

Keywords  Weibo; Users Relationship; Weibo Content;  User Modeling; 

 Personalized Recommendation;  User Profilling

本 科 毕 业 论   文 第 I 页

目 次

1 引言 1

1。1 研究背景 1

1。2 研究意义 1

1。3 研究现状 2

1。4 章节安排 4

2 文献综述