[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]

Journal of Information Science and Engineering, Vol. 37 No. 1, pp. 243-257

Building Student Course Performance Prediction Model Based on Deep Learning

1Department of Computer Science and Information Engineering
National Taipei University of Technology
Taipei, 106 Taiwan
E-mail: jykuo@ntut.edu.tw; colin19940702@gmail.com

2Institute for Information Industry
Taipei, 106 Taiwan
E-mail: pfwang@iii.org.tw

3School of Biomedical Engineering
Health Science Center
Shenzhen University
Shenzhen, 518060 P.R. China
E-mail: leiby@szu.edu.cn

The deferral of graduation rate in Taiwan’s universities is estimated 16%, which will affect the scheduling of school resources. Therefore, if we can expect to take notice of students' academic performance and provide guidance to students who cannot pass the threshold as expected, the waste of school resources can effectively be reduced. In this research, the recent years' student data and course results are used as training data to construct student performance prediction models. The K-Means algorithm was used to classify all courses from the freshman to the senior. The related courses will be grouped in the same cluster, which will more likely to find similar features and improve the accuracy of the prediction. Then, this study constructs independent neural networks for each course according to the different academic year. Each model will be pre-trained by using Denoising Auto-encoder. After pre-training, the corresponding structure and weights are taken as the initial value of the neural network model. Each neural network is treated as a base predictor. All predictors will be integrated into an Ensemble predictor according to different years' weights to predict the current student's course performance. As the students finish the course at the end of each semester, the prediction model will continue track and update to enhance model accuracy through online learning.

Keywords: deep learning, neural network, denoising auto-encoder, ensemble learning, prediction model

  Retrieve PDF document (JISE_202101_15.pdf)