Research On Early Warning Model Of Students’ Academic Work Based On Educational Data Mining

Colleges and universities use the student academic early warning model to enable administrators to detect students with learning difficulties early,so that they can be reminded and intervened in a targeted manner to help students respond early to avoid problems such as inability to complete their studies.This paper takes improving the performance of the student’s academic early warning model as the research goal,focuses on data preprocessing and model establishment,and improves the problems in feature selection and missing value processing,and proposes a student academic early warning model based on the PSO-XGBoost algorithm.The main work in the article includes:First of all,according to the needs of student performance early warning data set mining,the network public data set and actual data set were investigated and collected.Three sets of student performance data sets were obtained.Through analysis and comparison,a set of public data sets and one set were finally selected.The actual data set is used as the mining object of this paper,and the data is preprocessed.Then,based on the PSO algorithm and the XGBoost algorithm,a student academic early warning model was established.The experimental results were verified using the data set,and it was found that the XGBoost tree model can fill in the missing values in the data set to improve the feature selection ability of the model;and the individual memory characteristics of the particle swarm algorithm can select the best feature vector.Realize the fast convergence of the algorithm.For this reason,this paper proposes a student academic early warning model of PSOXGBoost algorithm that combines the advantages of the above two algorithms,and uses the algorithm and the mining performance of a single algorithm in the data set of this paper to verify the effectiveness of this model.The results show that the experimental accuracy of the academic performance early warning model based on PSO-XGBoost is higher than that of the academic performance early warning model based on the PSO algorithm and the academic performance early warning model based on the XGBoost algorithm.