MOOC环境下领域知识点的多重滤波提取  

Multiple filter extraction of domain knowledge points in MOOC environment

作　　者：陈之翼 王聪 李敏[1,3] CHEN Zhi-yi;WANG Cong;LI Min(College of Computer Science,Sichuan Normal University,Chengdu 610101,China;College of Movie and Media,Sichuan Normal University,Chengdu 610068,China;Network and Data Security Key Laboratory of Sichuan Province,University of Electronic Science and Technology of China,Chengdu 610054,China)

机构地区：[1]四川师范大学计算机科学学院,四川成都610101 [2]四川师范大学影视与传媒学院,四川成都610068 [3]电子科技大学网络与数据安全四川省重点实验室,四川成都610054

出　　处：《计算机工程与设计》2020年第12期3425-3431,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(61602331);四川省重点实验室开放课题基金项目(NDSMS201606);四川省教育厅重点基金项目(17ZA0322);四川省教育厅科研基金项目(17ZB0361)。

摘　　要：为解决智慧教育环境下的领域知识点提取问题,以MOOCs环境为基准,提出一种面向英文视频课程的领域知识点提取方法。利用词法和浅层句法分析提取字幕文本中的名词实体,借助参考文本分别对名词实体进行交集、词频阈值及TF-IDF过滤,去除噪声,生成候选知识点。借助公开的领域词典进行实体匹配,将匹配成功的实体标记为领域知识点。在此基础上提出基于结构化语义知识库倒向实体迭代相似度算法对未匹配词进行类别归属判断,识别词组形式的领域知识点。通过对真实数据集的测试,验证了算法的有效性。To solve the problem of domain knowledge point extraction in the context of smart education,based on the MOOCs environment,a domain knowledge point extraction method for English video courses was proposed.Lexical and shallow syntactic analysis was used to extract the noun entities in the subtitle text,and the reference text was used to perform intersection operation,word frequency threshold and TF-IDF filtering on the noun entities respectively to remove noise and generate candidate knowledge points.Entity matching was marked as a domain knowledge point by entity matching with a public domain dictionary.On this basis,a backward entity iterative similarity algorithm based on structured semantic knowledge base was proposed to classify the unmatched words to identify the domain knowledge points in the form of phrases.The effectiveness of the algorithm is verified by testing the real data set.

关 键 词：智慧教育 知识点提取 名词实体 去除噪声 相似度算法 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]