A Teaching Analysis System Based On Knowledge Graph And Text Analysis

In the context of education reform,more and more eyes are focused on basic education.However,the traditional classroom teacher teaching mode has been difficult to meet the current three-party interaction needs of parents,teachers,and students.In order to improve the above pain points,we designed and initially implemented an automated teaching analysis and evaluation system framework,which is mainly composed of a text analysis module,a knowledge point processing module,and a final comprehensive analysis module.Classroom text data belongs to short text,and it is difficult to classify short text such as sparse features,irregular language,and new words.We propose a multi-feature fusion short text classification method suitable for classroom teaching texts.Based on the traditional feature extraction of short-text characters,this method adds event-level information and expands conceptual information with the help of external databases.This method has achieved 55.2% and 77.5% classification accuracy on the Weibo sentiment7 classification data set and the e-commerce review 2 classification data set,which are also short text,leading the current cutting-edge methods.In addition,the classification accuracy rate of 95.7% was obtained on the SOUGOU news data set that is a regular text.This method has achieved a classification accuracy of 92.3% on the self-made classroom teaching data set,leading the current mainstream method.In addition,we also propose a set of knowledge point processing processes for classroom text,including data pre-processing,entity recognition,relationship extraction,knowledge graph construction and other processes,research on extracting knowledge points from text information,and combing intrinsic associations,comparing teachers The teaching plan and the syllabus issued by the Educational Examination Institute determine the coverage of knowledge points,the consistency of the teaching and the teaching plan,and the focus and depth of the class content.At the same time,the knowledge graph is built based on the neo4 j database,which is convenient for students to review.Finally,we propose a comprehensive analysis and evaluation system.With the help of text scene classification results and timeline information,we designed several classroom indicators,which are helpful for identifying the diversity of class formats,structural processes,and activity.It can help teachers review the teaching process,conduct benchmarking analysis.