基于典型度的相对熵相似度知识推荐

发布时间：2018-03-29 00:13

本文选题：互联网学习　切入点：知识推荐　出处：《天津大学》2016年硕士论文

【摘要】：互联网的发展加速了信息传递,不仅为教学提供了丰富教学资源,互联网在线学习模式也逐渐发展起来。不同于传统学习模式,互联网在线学习模式能够为用户提供不受时间,空间约束等特点,且丰富的互联网知识资源能够激发学生自主性,让在线学习者在学习过程中更有兴趣。但是,互联网迎来大数据时代同时,在线学习系统面临着海量在线学习资源与在线学习个体个性化需求难以匹配的难题。论文主要工作有:(1)将在线学习用户在线学习过程类比于人的认知过程,利用认知发展理论思想建立用户与知识资源关系矩阵。(2)提出相对熵相似度,并结合基于典型度推荐算法,提出基于典型度相对熵推荐算法。(3)将基于典型度相对熵推荐算法引入到知识推荐中,提出基于典型度相对熵在线知识推荐算法框架,为用户的在线学习提供个性化服务。论文首次提出基于典型度相对熵相似度推荐算法,并将基于典型度相对熵相似度应用到互联网知识推荐中,并实验证明了基于典型度相对熵推荐算法的正确性与有效性。基于典型度相对熵推荐算法有利于学习者的互联网在线学习,为未来建设更加智能化,个性化互联网学习平台做贡献。
[Abstract]:The development of the Internet accelerates the transmission of information, which not only provides rich teaching resources for teaching, but also develops the online learning model of the Internet. Unlike the traditional learning model, the online learning model of the Internet can provide users with no time. Space constraints and other characteristics, and abundant Internet knowledge resources can stimulate student autonomy, so that online learners are more interested in the learning process. However, the Internet ushered in the era of big data at the same time, The online learning system is faced with the difficulty of matching the huge amount of online learning resources with the individual personalized needs of online learning. The main work of this paper is to compare the online learning process of online learning users to the cognitive process of people. Using cognitive development theory to establish the relation matrix of user and knowledge resource. (2) the relative entropy similarity is proposed, and the recommendation algorithm based on canonical degree is proposed. The recommendation algorithm based on the relative entropy of canonical degree is introduced into the knowledge recommendation, and the online knowledge recommendation algorithm based on the relative entropy of the degree of canonical degree is proposed. In this paper, the relative entropy similarity recommendation algorithm based on canonical degree is proposed for the first time, and it is applied to Internet knowledge recommendation based on the relative entropy similarity of canonical degree. The experiment proves the correctness and validity of the recommendation algorithm based on the relative entropy of canonical degree, which is beneficial to the online learning of learners, and is more intelligent for the future. Personalized Internet learning platform to contribute.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.3

【相似文献】