A Classification Method Based Deep Learning for Educational News Comments

This article designs a classification method for long educational news texts based on labeled educational news reports. Firstly, CNN is used to extract local features of text information. Then, a bidirectional long short-term memory network is used to extract contextual information from the CNN output vector, thereby achieving effective extraction of text feature information. Secondly, in order to enhance the feature expression of the text, an attention layer is added at the end of the model. The model obtained through fusion not only solves the problem of different words in the text having different impacts on the classification results. Finally, the improved classification efficiency of this method was verified through education news review data.