We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Predicting high-risk students using Internet access logs.
- Authors
Zhou, Qing; Quan, Wenjun; Zhong, Yu; Xiao, Wei; Mou, Chao; Wang, Yong
- Abstract
Predicting student performance (PSP) is of great use from an educational perspective, especially for high-risk students who need timely help to complete their studies. Previous PSP studies construct prediction models mainly on data collected from questionnaires or some specific learning systems. Instead, students’ Internet access logs were used in this study to predict high-risk students. Since the raw data in log files are high-dimensional, complex and full of noise, several methods were proposed for the preprocessing of the data source. A high-dimensional feature selection framework is then designed to prepare features for the construction of a prediction model with good trade-off between computational efficiency and prediction performance. Experiments showed that the proposed prediction model can identify about 85% of high-risk students. Some online characteristics of high-risk students were also discovered, which might help student counselors and educational researchers better understand the relationship between students’ Internet use and their academic performance.
- Subjects
INTERNET in education; INTERNET access; ONLINE education; AT-risk students; PREDICTION models
- Publication
Knowledge & Information Systems, 2018, Vol 55, Issue 2, p393
- ISSN
0219-1377
- Publication type
Article
- DOI
10.1007/s10115-017-1086-5