社交网络中教育资源推荐的目标用户挖掘研究

发布时间：2018-01-09 02:15

本文关键词：社交网络中教育资源推荐的目标用户挖掘研究　出处：《中央民族大学》2017年硕士论文　论文类型：学位论文

【摘要】：互联网和信息技术正处于飞跃发展时期,随之产生的网络信息内容日渐丰富,网络数据也日渐增加。在这个互联网和教育并行发展的时代,人们更乐于借助社交网络平台寻求更多获取教育资源的途径,然而在有限的时间内准确快速获取需要的教育资源成为研究的重点。于是,数据挖掘技术、信息推荐技术、分类技术等前言技术应运而生。新浪微博社交网络平台提供了一个全面的分析用户兴趣的庞大数据源,成为近几年研究的热点。如何有效地给用户推荐有用的信息,需要找到关注某一资源或主题的目标用户,其中重要的研究工作即如何能准确分析目标用户的兴趣偏向。论文的主要研究内容如下:(1)研究社交网络目标用户数据采集相关技术和预处理的过程。其中,数据采集主要是利用基于Scrapy的网络爬虫技术,得到用于实验分析的用户基本属性信息和博文信息。数据预处理包括对语料去除停用词、分词等。(2)基于内容的用户兴趣特征提取与目标用户挖掘研究。利用LDA模型进行主题建模提取用户特征,将直接使用LDA建模和改进的LDA建模的实验数据进行对比分析,结果发现改进的LDA模型进行主题建模提取的特征主题准确性更高。选择基于聚类的半监督算法作为目标用户挖掘研究的分类算法,实验结果显示分类结果能较准确表示用户的关注倾向。(3)系统可视化技术和实现的研究。主要通过Java后端数据的处理和HTML CSS技术的前端展示实现,将目标用户分类结果界面化显示。文章研究的最终目的是为小学阶段学生以及关注小学教育资源的学生家长服务,系统能快速准确获取相关教学资源。同时,更有助于学校教育管理者开展工作,利用该系统可以将教学资源准确且有针对性的推荐给目标用户。
[Abstract]:Internet and information technology are in a period of rapid development, resulting in the increasingly rich content of network information, network data is also increasing. In this era of parallel development of Internet and education. People are more willing to seek more access to educational resources with the help of social network platform. However, accurate and rapid access to educational resources in a limited time has become the focus of research. Therefore, data mining technology. Information recommendation technology, classification technology and other preface technology came into being. Sina Weibo social network platform provides a comprehensive analysis of user interests of a huge data source. It has become a hot topic in recent years. How to effectively recommend useful information to users needs to find the target users who pay attention to a certain resource or topic. The important research work is how to accurately analyze the interest bias of the target users. Research on the social network target user data acquisition technology and preprocessing process. Data acquisition is mainly based on Scrapy based web crawler technology to obtain user basic attribute information and blog information for experimental analysis. Data preprocessing includes the removal of discontinuation words from corpus. Segmentation, etc.) based on the content of user interest feature extraction and target user mining research. Using LDA model for topic modeling to extract user features. The experimental data of direct use of LDA modeling and improved LDA modeling will be compared and analyzed. The results show that the improved LDA model is more accurate in feature topic extraction. The clustering based semi-supervised algorithm is selected as the classification algorithm of target user mining research. The experimental results show that the classification results can accurately indicate the user's concern. System visualization technology and implementation research. Mainly through the Java back-end data processing and HTML CSS technology front-end display realization. The final purpose of this paper is to provide services for primary school students and parents concerned about primary education resources, and the system can quickly and accurately obtain relevant teaching resources. It is also helpful for school education administrators to carry out their work, and the system can be used to recommend the teaching resources to the target users accurately and pertinently.
【学位授予单位】：中央民族大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13;G434

【参考文献】