Empirical analysis of university online public opinion based on the BERT-LSTM model
With the rapid development of social media,the impact of online public opinion on the image of universities and the psychological well-being of students is becoming increasingly prominent.Misinterpretation of information can lead to biases and doubts among students,potentially damaging the reputation of universities and undermining efforts in patriotic education.Therefore,this study aims to construct a supervised semantic analysis model to identify negative public opinion about universities on the internet,allowing for timely measures to be taken in response.Firstly data is collected by scraping the hot search page of Sina Weibo,followed by data cleaning.Subsequently,a sentiment analysis model is trained using BERT pre-trained models and bidirectional LSTM classification models to classify public opinion text.Secondly,the effectiveness and stability of the model are verified through the calculation of model evaluation metrics.Finally,the practical effectiveness of the classification model in identifying and responding to negative public opinion about universities is demonstrated through several examples.This model aids university administrators in promptly perceiving public opinion trends,swiftly restoring the truth,stabilizing campus discourse,and maintaining a positive image of the university and the psychological well-being of students.
NETL
NSTL
万方数据
