社会学习网络的分类方法研究与设计

发布时间：2019-01-26 18:52

【摘要】：随着互联网技术的发展,当今社会早已进入信息爆炸的时代,人们日益增长的知识需求也已经超出传统教育模式所能满足的范围。如何在信息世界里充分利用信息资源,营造个性化学习环境,满足人们随时随地学习的需求,成了当务之急。社会学习网络是基于Web2.0技术和大数据挖掘技术所构建的用于知识发掘、整合、存储以及传播的网络,能够根据用户信息提供个性化学习方案,并通过文本、语音、视频等人机交互方式将知识提供给用户,满足人们随时随地学习的需求。社会学习网络虽然能够满足人们在信息爆炸时代的学习需求,但是由于仍处于发展初期,其智能决策机制还远不够完善。论文基于移动互联网和大数据挖掘技术的发展,研究分类方法在社会学习网络中的设计与应用,主要研究工作和创新点包括以下方面:1)为了获取真实可靠的研究数据,论文首先设计开发了一个在线社会学习互联网平台。与其他在线学习平台不同,该平台不仅提供了课程讨论、教学视频播放等学习功能,还提供信息引导、校园互动等社交功能,为社会学习网络的平台设计与实现提供了参考方案。作为分类方法研究的平台基础,该平台实现了社会学习网络的数据采集功能。2)社会学习网络的课程讨论区用于用户之间的交流,但目前存在着信息泛滥、杂乱的现象,给用户带来极大的不便。针对这一问题,论文利用数据挖掘技术,提出了一种课程讨论区信息分类与排序方法,帮助用户快速获取有用信息。与单门课程分类方法不同,论文利用n门课程的主题信息组成数据源,将二元不平衡分类问题转换成n+l元平衡分类问题,从而提高课程主题分类与排序的准确性。实验结果验证了所提主题分类与排序算法的有效性,完善了社会学习网络的智能机制。3)社会学习网络保存了大量的用户视频点击信息,但这些信息都没有得到利用,造成极大的信息浪费。针对这一问题,论文提出了一种基于视频点击流数据的用户分类方法,利用用户点击行为来识别用户本身的学习程度。与其他视频点击行为研究不同,论文基于在线学习互联网平台改进用户点击事件设计,建立新的用户点击模型,并将用户点击行为与学习程度联系起来。实验结果验证了所提视频学习用户分类算法的准确性,完善了社会学习网络的个性化策略。
[Abstract]:With the development of Internet technology, the society has entered the era of information explosion, and the increasing demand for knowledge has already exceeded the scope of traditional education model. How to make full use of information resources in the information world, to create a personalized learning environment, and to meet the needs of people learning at any time and anywhere has become an urgent task. Social learning network is based on Web2.0 technology and big data mining technology for knowledge mining, integration, storage and dissemination of the network, can provide personalized learning programs according to user information, and through text, voice, Human-computer interaction, such as video, provides users with knowledge to meet the needs of learning anytime and anywhere. Although the social learning network can meet the learning needs of people in the era of information explosion, its intelligent decision-making mechanism is far from perfect because it is still in the early stage of development. Based on the development of mobile Internet and big data mining technology, this paper studies the design and application of classification method in social learning network. The main research work and innovations include the following aspects: 1) in order to obtain real and reliable research data, This paper first designs and develops an online social learning Internet platform. Unlike other online learning platforms, the platform not only provides learning functions such as course discussion, instructional video playback, but also provides social functions such as information guidance, campus interaction, etc. It provides a reference scheme for the design and implementation of social learning network platform. As the basis of classification research, the platform realizes the data collection function of social learning network. 2) the course discussion area of social learning network is used for the communication between users. It brings great inconvenience to users. In order to solve this problem, this paper proposes a method of information classification and sorting in course discussion area by using data mining technology, which can help users to obtain useful information quickly. Different from the single course classification method, the paper uses the topic information of n courses to form a data source, and converts the binary unbalanced classification problem into the n-l element balanced classification problem, thus improving the accuracy of course topic classification and sorting. The experimental results verify the validity of the proposed algorithm and improve the intelligent mechanism of social learning network. 3) Social learning network preserves a large number of user video click information, but these information are not utilized. Cause a great waste of information. In order to solve this problem, a user classification method based on video click-stream data is proposed in this paper. The user click-behavior is used to identify the user's learning degree. Different from other video click-behavior research, this paper improves user click event design based on online learning Internet platform, establishes a new user click model, and links user click behavior with learning level. The experimental results verify the accuracy of the proposed video learning user classification algorithm and improve the personalized strategy of social learning network.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP393.4;TP311.13

【参考文献】