Visual Analysis And Research Of Academic Performance Based On Student Behavior

With the advent of the era of big data,a large amount of education-related data has been generated,and these data have become an important force to promote innovation in education.Exploring the relationship between student behavior and academic perfor-mance is an important research direction in educational data analysis.Visual analysis is a method that uses human visual perception to analyze data through interactive exploration in recent years.It can integrate human cognitive abilities in the analysis of large-scale,high-dimensional and information-sparse educational data.Compared with the traditional method relying only on computer automation analysis,It can take advantage of people's cognitive abilities and more effectively explore the inherent patterns in data.However,in this new discipline,the existing visual analysis of student behavior and performance changes is mostly based on learning behavior data generated during the stu-dents' online course learning.It lacks research work to explore students 'life patterns at school and explore the relationship between students' life behavior and academic perfor-mance.In this paper,we explore the connection between student behavior and performance changes from the perspective of student life at school based on large-scale campus card data of student swiping behaviors.Due to the large scale and time-sequential character-istics of behavioral data,avoiding visual confusion due to the large volume of data has become one of the difficult points in this study.In addition,the behavioral data has a low information density.In this study,it is necessary to extract important characteristics from the behavioral data to analyze the relationship between behavior and performance.In this paper,a visual analysis framework combined with a predictive model is designed,and the visual analysis system EduRedar is implemented to study students' daily behavior pat-terns and explore the relationship between student behavior and performance changes?The main research work in this paper is as follows:(1)Based on the theory of visual analysis,a visual analysis framework combining pre-dictive models is proposed,and the automated analysis capabilities of predictive models are used to assist visual analysis.(2)Establish a performance prediction model in the visual analysis framework,and refine the importance of features through the prediction model,so that performance data is associated with behavior data.(3)This article analyzes the goals of visual analysis from the investigation of school staff's daily work,and formulates the visual view design basis.Designed corresponding visual mapping,visual layout algorithms and interaction methods for high data dimensions and time series(4)Integrate the visual analysis view and design and implement the visual analysis system,EduRedar,which is based on the performance prediction model,and supports multi-dimensional and multi-level analysis of data.We use real campus card swipe behav-ior data to explore student behavior and performance changes.