Research On The Influencing Factors Of Primary And Middle School Students’ Academic Performance From The Perspective Of Big Data

Under the background of the rapid development of society,education has become the focus of attention from all walks of life.In recent years,our country has made great efforts to develop education and emphasizes the importance of students’ academic performance,and has taken academic performance as an important reference for school education.It is well-known that students’ academic performance is directly related to the level of education.Therefore,it is of great significance to study the influencing factors of academic performance from many aspects.Through the analysis of many influencing factors of academic performance,we can have a deeper understanding of students’ learning environment and conditions,and take effective measures to help students improve their academic performance.This study is based on the CFPS family tracking survey data in 2018,taking the student’s class performance grade(sequencing variables)and academic self-efficacy(quantitative variables)to reflect the comprehensive academic performance situation.From the aspects of students and families,118 variables are chosen,including the personal basic information,the education,self-evaluation,family background,family relationship and family involvement.Firstly,this paper reviews the research results on the influencing factors of academic performance.Secondly,the current situation and the relationship among the studied variables are investigated by descriptive statistical analysis and correlation analysis.Thirdly,the Lasso Logistic model and Light GBM classification model were established for the grades of academic performance.The results of Logistic and Light GBM were compared and analyzed to get the important influencing factors of academic performance.Last,the quantile regression model based on Lasso variable screening method was established for academic efficacy to obtain the influence of various factors.The results showed that the influence of individual factors was more prominent in Logistic regression,while the family factors were more prominent in Light GBM.In Logistic regression,conscientiousness,openness and emotional instability had significant effects on academic efficacy and academic ranking,respectively.It is found that appropriately extending weekend study time and shortening non-weekend study time can effectively improve performance and academic efficacy.In Light GBM results,family factors such as parents’ total income from work,parents’ understanding of their children’s whereabouts,parent-child relationship,parent-child meeting frequency,injury cost and after-school tutoring are more important.From the perspective of model classification effect,Light GBM model is better than Lasso Logistic model,with an accuracy of 0.9899.Deviant behavior has a significant negative impact on academic ranking and academic efficacy,and the number of students’ pocket money has a positive impact on Logistic regression and Light GBM.In conclusion,students’ non-cognitive ability,parent-child relationship and various economic input factors have a greater impact on academic performance,and society and parents should pay more attention to family economic background,parent-child relationship and non-cognitive ability cultivation environment.