Research On Predicting Academic Procrastination In Online Learning

With the development of educational data mining and artificial intelligence,the use of online learning behavior data to evaluate students’ performance has become a trend.Educational data mining technology is an important technical tool to support the development of online teaching research.And it has greatly promoted the reform of education and teaching.Educational data mining received extensive attention and research exploration from scholars at home and abroad.However,relatively few studies have used educational data mining technology to diagnose and predict the phenomenon of academic procrastination in the online learning process.Academic procrastination is one of the important behavioral manifestations in the process of online learning.Academic procrastination often has a negative impact on students’ online learning effectiveness.Therefore,in order to improve students’ online learning effectiveness and academic performance,this thesis effectively predicts students’ academic procrastination tendency and dynamically diagnoses students’ procrastination,so as to intervene and help correct it in time.In this thesis,the online learning behavior log data of college students at a university were used to construct an academic procrastination prediction model and conduct academic prediction research on online learning.(1)Firstly,we constructed an online learning academic procrastination model based on three dimensions: time management,self-regulation,and study readiness.We selected 10 online learning behavior variables as the characteristic variables for analyzing students’ academic procrastination.Three categories of students were obtained by cluster analysis method: severely procrastinating students,average procrastinating students and active learning students.In this thesis,the close relationship between academic procrastination and academic performance was analyzed,and students’ procrastination behavior was predicted based on the academic procrastination model.It was found that the characteristics of student groups with different levels of procrastination differed significantly.With severely procrastinating students procrastinating in completing academic tasks and performing poorly academically,and actively learning students having high learning interaction and outstanding performance.The prediction model of academic procrastination based on comprehensive indicators constructed in this thesis can predict students’ tendency to procrastinate more effectively.(2)Based on previous studies related to academic procrastination and the above process of predicting academic procrastination based on comprehensive indicators,it is found that good time management skills are important for improving students’ procrastination.And some online learning platforms cannot exhaustively collect multi-dimensional comprehensive indicators.In view of the above reality,this thesis constructs an online learning academic procrastination prediction model based on time management from the perspective of students’ time management.Using the characteristics of the engineering method to online learning task completion time is expressed as structured characteristic vector.One-Hot encoding of students’ task completion times was used to effectively judge students’ procrastination tendency based on their time management ability.The K-means algorithm was used to analyze students’ procrastination performance,label students as procrastinating and non-procrastinating students.Eight prediction methods were used to predict academic procrastination,and the performance of different prediction methods was evaluated.The thesis compensates for the lack of data acquisition in some online education platforms and makes the prediction process more universal.In addition,it is found that the academic procrastination prediction model based on online learning task completion time has good performance in predicting students’ procrastination,and the prediction accuracy is mostly above90%.It is expected that the research results of this thesis can provide some warning effect for online learners,improve the phenomenon of student procrastination,and motivate students to maintain their motivation and initiative to improve their learning performance.