| Knowledge Building and its platform Knowledge Forum(KF)are the most representative knowledge innovation theory and technology in the international learning sciences community,which represent the direction of teaching and learning reformation in the 21 st century.The theory is to construct knowledge at multiple levels,such as individual,group and community,which focuses on “idea” and its evolution process.Knowledge Forum accumulates large and complex semi-structured data over time.In the actual knowledge building environment,both students and teachers have to deal with these large amounts of data in a short period of time,and then diagnose the learning process of students and predict their further improvement.The original functions in Knowledge Forum are too simple to carry out more complex and in-depth semantic analysis.Educational knowledge graphs are a kind of knowledge network formed by knowledge entities and their semantic connections.The educational knowledge entities stand for both individual knowledge structure and collective intelligence.As an advanced way of knowledge organization,educational knowledge graphs are an effective method to correlate domain knowledge entities,which can improve the data analysis and management function in Knowledge Forum to a large extent,thus generate the cognitive schema of learners and aggregate online education resources effectively.This study is to utilize data mining and artificial intelligence technology to handle the text data in Knowledge Forum.The model of constructing educational knowledge graphs is put forward,mainly including entity extraction,entity relation extraction,normalized graph storage,graph visualization system implementation,which can enhance the analysis function and support the knowledge visualization.The main research work and innovation of this paper include the following points:(1)In order to improve the accuracy of entity extraction in Knowledge Forum and reduce the dependence on the size of dataset,an entity extraction algorithm based on statistics and topics is proposed.Firstly,we improve the accuracy of Chinese word segmentation by mutual information and linguistic rules,then obtain entities through the improved TF-IDF algorithm based on dictionary and location.Secondly,we get deep semantic information through topic model LDA,and then extract topic entities.Finally,Word2 vec tool is used to calculate the word vectors' cosine values,which can be used to gain the relevant entities and expand the size of entity dataset.(2)We proposed the extraction algorithm of educational entity relationship oriented to Knowledge Forum,which can get the association relationship and hierarchical structure among entities.For one thing,according to different support and confidence coefficients,we can extract the association relation using the classic association rules algorithm Apriori.For another,the hierarchical relation is extracted by using the pattern matching method on the basis of the characteristics of Chinese words.The experimental results show that the accuracy and coverage of entity extraction and entity relationship extraction are superior,and the constructed educational knowledge graphs are highly readable,which are very suitable for the knowledge organization and reflection,thus lay a solid foundation for the recommendation and application of the teaching resources in Knowledge Forum.(3)For the sake of helping teachers and students have access to the educational knowledge graphs in Knowledge Forum intuitively and quickly,we store the educational knowledge graphs through the graph database Neo4 j and design the node type,attribute type and relationship type.Bootstrap and SSM are used to develop a visual system of educational knowledge graphs for Knowledge Forum,which have the functions of intelligent retrieval and data management.In addition,Echarts provides powerful supports for dynamic visualization function of the graph system. |
PDF Full Text Request