Design and Construction of Network Public Opinion Analysis System based on Certain Topic

摘要

With the rapid development of the Internet, the Internet has provided a broad platform for more and more people to express their emotions and opinions. The network has gradually developed into a gathering place for public opinion. In this background, negative, negative network public opinion may deceive and mislead the public, causing a growing threat to public safety to the society, it is necessary to carry out effective analysis on the network topic and speech, timely capture the dynamics of public opinion, has important practical significance to maintain social stability and the construction of a harmonious society. Therefore, the study of network public opinion analysis technology has become a very urgent and important issue. With the development and popularization of the Internet, it has become the fourth media after newspaper, radio and television ". More and more people are willing to share information on the Internet, to express their views and views on certain topics, forums; micro-blog has become an important carrier of social information dissemination. However, due to the low threshold and openness of the Internet has led to a certain extent, the network has become a public rumor breeding and dissemination of the land, to a certain extent, to stimulate the social contradictions and affect social stability. To effectively carry out the network public opinion monitoring has become an urgent. This paper points out the importance and necessity of network public opinion monitoring, introduces the domestic and foreign research present situation in related fields, and summarizes the related technologies and theories used in the system design and implementation. In order to solve the problem of network public opinion, a general solution framework is proposed to provide practical ideas for the design and implementation of similar systems. In the system analysis module and monitoring module two large public opinion public opinion "emotion based on the text as the core, with interactive interface, database and index files as the basis, given the data acquisition, database design, sentiment analysis, heat topic situation prediction and interface design method. The whole system can be independent of Web data crawling, sentiment analysis of the data, and by using the analysis results to predict heat topic trend, is a fully functional system; correction of the existing heat topic evaluation method, put forward a new model for predicting the heat topic trend, and realize the heat topic situation forecast in the system. The existing topic evaluation index simply considers the number of clicks and replies, without taking into account the impact of the point of view on the topic; the prediction model of the topic situation is not consistent with the actual situation. In this paper, the point of view of the tendency factor into the existing heat calculation formula, and put forward the use of Gama model to predict the topic of the heat situation, and finally to the idea of verification and implementation. Experimental results show that the improved method is more in line with the actual situation. How to enhance the detection ability of Internet public opinion, which is the direction of the government, needs to pay attention. In this paper, the Windows based platform, Apache server and combined with the current advanced Chinese Natural Language Processing technology, construct a highly efficient, accurate and practical network public opinion monitoring system. This system uses the JAVA language development; the server uses SQL Server MS, and combined with the AJAX technology to build the page display technology. Network public opinion, such as the hidden nature of the characteristics of a wide range of related technologies (Chinese word segmentation technology, text mining technology) applications, but also on the network public opinion monitoring put forward higher requirements. This paper focuses on the design and implement