Construction And Application Of Academic Early Warning Model Based On Cognitive Diagnosis

With the application of big data,artificial intelligence,cloud computing and other new technologies in the field of education,it promotes the deep integration of education informatization,helps the realization of personalized learning,accurate management and intelligent services,strengthens the talent cultivation concept of ability first,and promotes the innovation of intelligent education.It not only provides strong technical support in obtaining students’ learning process data,but also provides effective support It is possible to analyze student process data and predict crisis students.The core of learning early-warning is based on scientific theory,and learning analysis technology to comprehensively evaluate learners’ multiple dimensions,such as knowledge,learning behavior,emotion and psychology,and according to certain standards to early-warning their future learning performance,send early-warning information or provide intervention measures,and feed back to learners,teachers and teachers To cultivate managers,help learners personalized learning,teachers’ accurate teaching and managers’ scientific decision-making.Therefore,how to use big data technology for learning early warning has become an inevitable trend.In recent years,more and more researchers are devoted to the study of learning early warning.However,the existing learning early warning research still has some defects:(1)most of the existing research on learning early warning is applied to online learning and university environment,which is for University learners.It is not applicable in the basic education environment and does not conform to the scene of basic education,but there are few research on learning early warning in the basic education environment;(2)most of the existing learning early warning models are for online learning and university environment For students’ learning style,it ignores the influence of students’ cognitive state on their studies,such as knowledge state,ability level,etc.(3)the existing learning early warning system only stays in the shallow application,but there is little research on how to affect students’ learning effect.Therefore,how to take root in basic education,according to the characteristics of traditional classroom teaching,comprehensively evaluate learners,and realize multi-dimensional and deep learning early warning for learners is the research focus of this paper.Therefore,the main contents of this study include the following aspects:First of all,by investigating the relevant literature of learning early warning,this study finds that there are still some problems in the study of learning early warning,such as the lack of learning early warning for basic education and traditional classroom,and puts forward new ideas and methods to solve them.A learning early warning model based on dual channel cognitive diagnosis method is constructed.According to the theory of cognitive diagnosis,using external data such as learners’ test scores,combined with DINA and Rasch model,the potential characteristics and cognitive structure of learners are jointly modeled,and a dual channel diagnosis model oriented to knowledge and ability is constructed;based on the external performance of learners,such as grades and rankings,and the internal characteristics of comprehensive ability and cognitive structure,this paper puts forward a new idea of learning early warning including knowledge point warning,fluctuation warning,wrong topic warning and learning intervention,and establishes a learning early warning model based on potential characteristics and cognitive structure;the effect and performance of the learning early-warning model based on cognitive diagnosis were compared and analyzed with other early-warning models as the baseline.Secondly,this research develops and implements a learning early warning system.In view of the learning early warning model based on cognitive diagnosis method constructed above,based on the needs of users and the system,around the learning early warning model based on cognitive diagnosis,this paper introduces the five functional modules of the learning early warning system,including login,testing,diagnosis,early warning and learning intervention,expounds the overall structure and architecture of the system,and develops and realizes the learning early warning system.Finally,with the help of learning early warning system,this study selects a middle school student in Suzhou City of Jiangsu Province as the experimental object,sets up experimental class and control class,tracks the students’ academic performance and change trend,and carries out a two-month empirical study.The results show that:(1)learning early warning system can help to improve the academic performance of learners,among which,compared with high and middle level learners,the improvement effect of low level learners is better;(2)learners with different academic levels have different perceptions of the system,such as high level students are more willing to express positive attitude than middle and low level students in cognitive diagnosis;low level students are more willing to express positive attitude in cognitive diagnosis Average students are more willing to express positive attitude than middle and high-level students in knowledge point early warning;(3)there is a positive correlation between learners’ perception of learning early warning system and learning strategies,learning strategies and academic performance,which means that the system can further improve learners’ academic performance by influencing learners’ learning strategies.