基于深度卷积神经网络的大学英语四级成绩早期预警

 The early warning models of CET-4 scores are easily disturbed by the differences in students daily behavior patterns, affecting prediction accuracy. A gray image database of students modular learning is established using the CET-4 modular learning scores on a platform, and deep learning is integrated with early warning to form an early warning model of CET-4 scores based on deep convolution neural networks. An early prediction was made on whether students can achieve the expected results in CET-4. The results show that the prediction model based on the deep convolution neural networks has higher accuracy than the existing prediction models and can get an earlier optimal intervention time. It benefits teachers intervention with relevant students and can help improve their CET4 scores and English language competency.