Deep learning and integrated learning for predicting student's withdrawal behavior in MOOC

Abstract

MOOC attracts students with its unique teaching mode and high-quality curriculum resources, but it also faces the problem of high dropout rate, which affects the long development of MOOC. In order to solve the problem of high dropout rate faced by MOOC platform, this paper proposes the method of combining deep learning and integrated learning to construct the prediction model of students' withdrawal behavior. The experimental data were collected from MOOCCube2020 dataset. The convolution neural network is used to extract hidden features from the original data, and the output features are used as the input of ensemble learning model. Then, various traditional classification methods are used for training and prediction, and the prediction results of various models are fused to obtain the final result. Experiments show that the model can well fit the correlation between students' learning performance and class quitting behavior, so as to accurately predict whether students will quit the course, which is helpful to the in-depth study of MOOC learning mode.