SOUA: Towards Intelligent Recommendation for Applying for Overseas Universities
Binqing Wu, Zhihao Ke, Min Fu, Yuwei Xia · 2019
In China, to pursue a master's degree in foreign countries is becoming an increasingly popular choice for undergraduate students. However, predicting the results of university applications is mostly done by applying empirical rules and referring to previous admission cases similar to the applicant's conditions, which is unreliable. In this paper, we propose an intelligent recommendation method, called SOUA, to make recommendations for students who apply for the admissions of overseas universities to study abroad. Our method is proposed based on a dataset obtained by us, which is about Chinese undergraduate students' applying for master's degrees in computer science-related majors of American universities. We apply data mining techniques on this dataset. A new data augmentation method named ILGA is proposed to expand the size of the training set. In the experiments, machine learning models such as LR, KNN, DT, SVM, RF, GBDT and MLP are investigated using this dataset. The experimental results show that RF achieving a weighed f1-score of 71.2% on the test set outperforms other considered machine learning models. And RF produces the optimal university recommendation accuracy of 80.9%.