摘要
通过个性化试题推荐,来对在线学习用户实现全面性的评估。个性化试题推荐的关键思想就是构建每个用户的知识点的知识图谱,即Skill-Graph,以通过挖掘丰富的历史试题成绩和网络中可用的对知识点掌握情况,全面建模那些在线学习用户的各项能力进行评估的能力。首先开发一种基于自适应门机制的双向LSTM-CRF神经网络的知识实体提取方法。接着为了提高提取的知识实体的可靠性,设计了一种基于实体-URL图上的标签传播方法,该实体-URL图是根据百度搜索引擎的查询日志中的点击数据构建的。此外,抽取知识实体之间的上位词-下位词关系,并通过利用具有广泛上下文特征训练的分类器来构建知识图谱。最后,提出了一种基于知识点的个性化试题推荐方法,以提高的用户学习效率和优越性。
Through personalized test question recommendation,a comprehensive evaluation of online learning users is achieved.The key idea of personalized test question recommendation is to build a knowledge graph of each user's knowledge point,namely Skill-Graph,to comprehensively model those online learning users by mining rich historical test scores and knowledge points available in the network.The ability to evaluate various capabilities.First,a knowledge entity extraction method based on the adaptive bidirectional LSTM-CRF neural network is developed.Then,in order to improve the reliability of the extracted knowledge entity,a label propagation method based on the entity-URL graph is designed.The entity-URL graph is constructed based on click data in the query log of Baidu search engine.In addition,extract the superordinate-subordinate relationship between knowledge entities,and construct a knowledge graph by using a classifier trained with a wide range of context features.Finally,a personalized test recommendation method based on knowledge points is proposed to improve user learning efficiency and effectiveness.
作者
王启亮
刘镇
WANG Qiliang;LIU Zhen(School of Computer,Jiangsu University of Science and Technology,Zhenjiang 212003)
出处
《计算机与数字工程》
2022年第5期1073-1077,共5页
Computer & Digital Engineering
关键词
知识点
数据挖掘
知识图谱
推荐系统
knowledge points
data mining
knowledge graph
recommendation system