BTCBMA Online Education Course Recommendation Algorithm Based on Learners' Learning Quality

To address the problems of existing online education curriculum recommendation methods such as low accuracy, an course algorithm (BTCBMA) considering learner learning quality is proposed. Firstly, BERT model combined with TextCNN to implement preliminary extraction text features. Secondly, convolution neural networks and BiLSTM are used capture deep features temporal in data. Finally, a multi-head attention mechanism extract key information from interaction sequences, review texts, multiple attributes. Experiments demonstrate that precision, recall, F1 values proposed method MOOC dataset 0.224, 0.241, 0.237, 0.239, respectively, while CN 0.217, 0.227, 0.233, performance significantly superior compared methods. For learners systems, can effectively recommend high-quality courses, which great significance for improving efficiency learners.