The Recommendation Module of College Students' Civic and Political Hot Spots by AM Strategy Incorporating Social Labeling

Current research on personalized Civic and Political education for college students is a hot topic, but most current recommendation algorithms or models are unable to achieve bidirectional recommendation selection between users and targets. In view of this, the study adopts convolutional neural network and attention mechanism to implement a hot recommendation model for college students' civics based on social label matching. The experimental results show that the average root-mean-square error value of the personalized recommendation model proposed in the study is 0.797 in the label generation task, which are lower than the average root-mean-square error values of the other models participating in the experiment, and thus the proposed model labels have an excellent label generation capability. In addition, in the label matching task, the average absolute errors of the proposed model on the two datasets are 0.704 and 0.706, which are much lower than those of the control models participating in the experiments. In the model performance test, the proposed model shows excellent performance, thus indicating that the model is state-of-the-art. The proposed model of the study contributes to the development of personalized civic education in universities and also plays a role in the reform of the education industry.