SA-FEM: Combined Feature Selection and Feature Fusion for Students’ Performance Prediction

Abstract

Around the world, the COVID-19 pandemic has created significant obstacles for education, driving people to discover workarounds to maintain education. Because of the excellent benefit of cheap-cost information distribution brought about by the advent of the Internet, some offline instructional activity started to go online in an effort to stop the spread of the disease. How to guarantee the quality of teaching and promote the steady progress of education has become more and more important. Currently, one of the ways to guarantee the quality of online learning is to use independent online learning behavior data to build learning performance predictors, which can provide real-time monitoring and feedback during the learning process. This method, however, ignores the internal correlation between e-learning behaviors. In contrast, the e-learning behavior classification model (EBC model) can reflect the internal correlation between learning behaviors. Therefore, this study proposes an online learning performance prediction model, SA-FEM, based on adaptive feature fusion and feature selection. The proposed method utilizes the relationship among features and fuses features according to the category that achieved better performance. Through the analysis of experimental results, the feature space mined by the fine-grained differential evolution algorithm and the adaptive fusion of features combined with the differential evolution algorithm can better support online learning performance prediction, and it is also verified that the adaptive feature fusion strategy based on the EBC model proposed in this paper outperforms the benchmark method.

E-Learning Behavior Categories and Influencing Factors of STEM Courses: A Case Study of the Open University Learning Analysis Dataset (OULAD)

“…Balti et al [32] categorize learning activities according to learning style theory, defining and analyzing behavioral traits in depth and thus revealing learners' learning preferences apart from the correlation between behavioral traits and learning achievement. Ye et al [33] reflect the existence of an intrinsic correlation between learning behaviors by constructing an E-learning behavior classification model (EBC), and the classification fusion of learning behaviors better supports E-learning prediction performance. Considering the student's dynamic cognitive structure during the learning process, Sun [34] divides the learners' E-learning behavior process into four stages: the learning occurrence stage, the knowledge acquisition stage, the interactive reflection stage, and the learning consolidation stage.…”