| Due to the anonymous technology,convenience and timeliness of the online community,it is easy to generate social public opinion,and the spread of wrong public opinion will also damage social security and endanger the public’s sense of security.Among them,public opinion events in the field of colleges and universities are the focus of national attention.Through the network community public opinion early warning technology that integrates multiple knowledge,it is possible to timely discover the current public opinions related to the field of colleges and universities on the Internet,help the government and relevant departments of colleges and universities to guide or intervene in public opinion,and help ensure public safety and social stability.In this paper,the idea of software engineering is used to construct a network community public opinion analysis and early warning system that integrates multiple knowledge.Firstly,the demand analysis of the system is carried out,including user analysis,functional and non-functional requirements analysis;Then a general design of the system is carried out,including the overall system design,functional module design,database design and interface design,and the design of core module;Next,the user management,data collection,public opinion analysis,public opinion early warning and other functional modules,the overall framework,and the core modules are implemented;Finally,the system test and performance analysis are carried out.This system completes the entire process of public opinion analysis and governance from collection to analysis to early warning,and integrates knowledge such as prior knowledge,sentiment analysis,semantic analysis,and algorithm model to build a public opinion analysis and early warning model of multi-knowledge fusion.After the collected online community’data is preprocessed,it is first filtered through the public opinion analysis model,including using abstract extraction technology to compress superlong texts,using neural network models to predict the emotional tendencies of texts,and using neural network models fused with prior knowledge to predict Text Sentiment Sensitivity.Then use the public opinion early warning model based on the network public opinion knowledge base to analyze the types and levels of early warning.Finally,it improves the recognition accuracy of public opinion in the field of network community colleges and universities,and provides support for the governance of public opinion in the field of colleges and universities. |
