Author(s)

Chen Hongxia, Jiang Lili

Corresponding Author

Jiang Lili

Abstract

Different groups of students have different levels of mastery of different English vocabulary. Based on the real teaching scenario of an Internet education company, we collected a large amount of student data for a certain English vocabulary learning course in the fifth grade of elementary school. After data preprocessing and correlation analysis remove weakly related attributes, a predictive model of the difficulty of mastering English vocabulary is constructed based on the Xgboost algorithm. By real data test, the predictive performance of the model is analyzed and discussed. Finally, the loss/profit ratio of the model based on the profit function and loss function are given and discussed.

Keywords

Psychological analysis, Cognitive rule, English education, Xgboost, Loss function