Intelligent Monitoring Platform Of Campus Network Public Opinion Based On Social Media

As an important part of the network environment,network public opinion has an important impact on network security.The occurrence of network public opinion in colleges and universities often contains the ideological views of most students.Therefore,the design of intelligent monitoring platform of public opinion based on the campus network environment can effectively help university managers to effectively analyze public opinion.Public opinion analysis is carried out from user comment data in social media,which involves the sentiment classification task of user comment text.In sentiment classification tasks,traditional methods require a lot of human work and low efficiency.With the development of related technologies,sentiment classification tasks based on deep learning also have certain room for improvement.At present,most sentiment classification tasks are binary or ternary classification tasks,and the fine granularity of sentiment classification results needs to be deepened.In view of the above shortcomings,this paper studies and implements an intelligent monitoring platform of campus network public opinion based on social media.The main work of this paper is as follows:1.Aiming at the problem of public opinion sentiment analysis through user comments,a method of batch obtaining user comments is proposed.By encapsulating the requested web page information,the method simulates the user accessing the target web page,parses the web page resources to obtain user comments and saves them to the local.Text denoising and text preprocessing were performed on user comment data,and the pre-trained ERNIE was used to complete the word vectorization of characters and words.2.Aiming at the problem of simple text sentiment classification results,this paper proposes an EBAF algorithm for sentiment feature extraction and classification,and realizes the classification of seven types of text sentiment.Firstly,the pre-trained language model was used to learn the complete semantic representation of the words in the sentence,the Bi LSTM and attention mechanism were combined to extract multi-dimensional emotional features,and the fully connected classifier was used to complete the mapping relationship between the emotional feature vector and the emotional label of the instance to determine the emotional attribute of the text.The EBAF algorithm achieves 87.7% precision,86.9% recall and 86.8% F1 score,which has better performance than other language models or feature extraction models.3.Complete the visual design of public opinion analysis for data display and interaction.The mapping relationship between data and charts is established,interactive charts are designed to realize the selection of data,and the key information contained in user comments is expressed in a variety of ways.The intelligent monitoring platform of campus network public opinion based on social media was implemented and tested.After the development of each module,the functional integrity of each module was verified by unit test,and the integrity of the platform was verified by integrating each module and testing the overall process.Through research experiments and platform implementation,the platform can better reflect the various functions of public opinion analysis,provide users with multi-dimensional data indicators under public opinion events,and part of the visual data has a certain degree of interactivity.The platform provides decision-making basis for the ideological guidance of students in the network environment of colleges and universities.The research results have certain theoretical significance and application value in the field of campus network public opinion monitoring and analysis.