Key Technology For Internet Public Opinion Analysis Of Massive Data Retrieval Model

In recent years, China’s Internet users have published a large number of self-views and comments on the Internet for social and public affairs, which constitute the prototype of the network public opinion. However, most of the Internet public opinion information is mostly insulated and distributed in nook and cranny of the network. Aiming at the characteristics of public opinion information, we need to use modern science and technology to design effective measures when we do research on network public opinion, and public opinion monitoring system came into being in this context.At present, Public opinion analysis system has a certain application in real life. The Government and relevant management departments will use public opinion monitoring and analysis system to make timely warning of public opinion and take measures to prevent the worsening of the situation. In this paper, after studying a large number of information about public opinion analyses at home and abroad, a simple Internet public opinion monitoring system is designed and implemented. In this paper, the work done as follows:(1) Researchin Query Expansion TechnologyIn the public opinion monitoring system designed in this paper, the user can enter the monitored keyword on the system query page. The systemmakesemantic extension for keywords timely, that is mainly based on the How Net semantic concept dictionary, calculatethe three wordsof the highest semantic similarity with keywords in semantic,used as candidate expansion terms.(2) Research in Information Retrieval TechnologyAfter the user submitsthe keywords, the scope and time frameneeded to monitor, the systemadopting vertical search technology to hierarchical system structure. Using the we b crawler for extendedkeywords, it will collect information from the monitoredweb sites; To the collection of information,using the general information extraction system for infor mation extraction and SCWS Chinese word segmentation, and proposed an improved feat ure word extraction algorithm, keyword location information is included in the account, a dding its weight calculation, good experimental results, Finally, the system calculate the d ocument similarity. (3)Applications of Public Opinion Analysis TechnologyAfter the above steps, process the document under text clustering and hot spots discovery algorithm, extracted the hot topic concluding more public opinion information, and the situation of public opinion to the outbreak of the region, and then after the text sentiment orientation analysis, makeemotional judgment forthe public opinion of Internet users through the manual annotation and How Net semantic dictionary. Finally, through automatic summarization technology formed allows users an intuitive understanding of the number of ground Digest Word reports and public opinion chart analysis results.