Application of Data Mining in Academic Early Warning

摘要

Nowadays, the phenomenon of college students failing in examinations is becoming more and more prominent. With the accumulation of failing credits, students lose the qualification to receive the degree certificate or repeat the grade, how to ensure that students complete their studies normally has become a hot topic in college education. This paper uses the test scores and student status information of the 2014 students as the data basis, analyzes the data, uses the Pearson correlation coefficient to study the correlation between the target attribute and other attributes, and uses the support vector machine, logistic regression, random forest, decision tree, and KNN five common classification algorithms to build Academic early warning models, uses Precision, ROC curve and cross-validation as model evaluation indicators. Experiments show that the recognition rate of the academic early warning model for students' academic performance is as high as 93.07%.