Application Of Topic Tracking And Visualization In The Agricultural Of Network Public Opinion

With the rapid development of Internet, more and more Internet information is showed to the users through the disorganized way. It is becoming more and more difficult for the users to detect, track and management the useful network information which they interested in. Especially, some burst events, the randomness and burst increase the difficulty of detecting and tracking, at the same time, the management cost gains. The agricultural related to National livelihoods, and it plays a crucial role in its social stability and economic development. In recent years, agriculture-related network events occur frequently, causing widespread concern in public opinion.The topic detecting and tracking technology uses the technology to organize the news by the topics, including finding the new topics, by this, it can track and report the topic. This article use the topic detecting and tracking technology to detect and track the agricultural events. Firstly, we collected a large number of training documents according to the uses’need. Secondly, we filtered a large number of testing documents. At last, we got the topic development situation, so that, we can master the topic’s panorama.This article main work including:1) Putting forward the topic model basing on the Topic and Term Frequency. Its basic idea is that the terms which appear in the topic frequently have more ability to express the topic. We ordered the quotient of Term Frequency of the topic and the number of document in the topic, and then got the first N quotient as the feature words.2) Adding the Burst Term based model to the1) model. The basic idea of the model is that the burst terms have more ability to express the burst story. So the Burst Term based model can reduce the Topic Shift situation.3) Selecting the seed events. The basic idea is that some stories of the topic can more or less express the topic. So using Agglomerative Clustering method finds some stories to represent the whole topic.4) Visualization by Fusioncharts. 5) Finding the popularity of the author and the popularity of the website. According to the information of authors, websites in the data base and the popularity Calculation methods, the hot authors and websites can be found.6) Crawling3867stories to do experiments.