Abstract Educational big data analytics can provide data support and decision-making reference for the optimization of education and teaching in colleges and universities by mining the daily behaviors of college and university students, searching for the data features that affect college students’ academic performance, and establishing corresponding prediction and early warning models. In this paper, we constructed a college student performance prediction model based on the CNNBiLSTM-Attention algorithm, integrating the advantages of bi-directional long- and short-term memory neural networks and attention mechanisms, introducing the 1-tanh function after the forgetting gate to make the output value of the forgetting gate in a more obvious range, retaining the features of the input data as much as possible, and improving the learning ability of LSTM. In the context of teaching in colleges and universities, we observed that the model in this paper achieved an accuracy of 86.39% and an F1 value of 83.59% on two dataset tasks, Principles of Microcomputer and News Editing, respectively. When predicting the results of the 2022 academic year grade performance points, the prediction accuracy of seven different score levels exceeded 80%. Three students in the experimental class had midterm scores below 40, but after teacher communication and targeted counseling, there were no students with scores below 40 at the end of the semester. In the midterm prediction, no student scored higher than 90. After the teachers adjusted the teaching plan and optimized the teaching arrangement based on the prediction results, six students had scores above 90 at the end of the final term. Students’ scores in all subsections have improved, and the overall distribution of scores has moved significantly toward the higher-scoring sections. From 68 points in the midterm prediction to 72 points in the final test, the number of sub-segments increased the most. The use of big data technology has had an optimizing effect on existing universities’ traditional education and teaching models.
Read full abstract