Objective In order to effectively monitor the rapidly growing online public opinion data of universities.Methods The effective monitoring and early warning of the university network public opinionwereachievedby ap-plyingthe optimized multi-channel BiGRU-CNN model with self-attention mechanism to sort university public sentiments into non-negative and negative ones,inwhich the self-attention mechanism of the BiGRU model-wasimprovedby usingthe position weight parameters,and the AdamW optimization algorithm was introduced to op-timize the entire model.Results Experiments showed that the BiGRU-CNN model had a comparatively higher per-formance,more effective in sorting university public sentiments.Conclusion The proposed BiGRU-CNN model can effectively monitor online public opinion data of universities and has good performance.
维普
万方数据