Process Platform Of Public Opinion For Short Message

With the e-commerce technology becoming more and more mature, andthe increasing popularity of mobile terminal------mobile phones, SMS (Short MessagingService) provides new tools for communication between the public, and it is accepted by more and more mobile phone users. As the SMS has been spreading widely day by day, the government departments have paid more attention to message public sentiment. Message public sentiment is different from the traditional one. In regard to message public sentiment, the occurrence range is wide, spreading speed is high, and their eruption spot are difficult to be detected and controlled. All above situation proved that: it is more important to effectively detect and control message public sentiment.Based on the analyzed of the domestic and international development of the hot-topic detect research and the text processing techniques, we deeply analyzed how to design an effective architecture of the SMS public opinion analysis system and what is the proper system process in this thesis, introduced the main function of the system management module, SMS information collection module, SMS information processing module and the hot topic detection module on the system. At the same time, we designed the database of the system.The paper analyzed the key technology that mentioned in the realization of the SMS Public Information Disposing Platform. It included the acquisition of SMS data, the preprocessing of text features, text clustering and hotspot discovery. After the deeply study of the key technology, the paper came up with a method of collecting data that is fit for SMS data, combined with the application of the key technology. By analyzing the SMPP protocol, the paper taken delivery of the SMS contents, stored them into the database, then made a Chinese dictionary separating words algorithm which was improved based on MM and RMM. The paper compared the similarity formula of the clustering method, selected the generalized Jaccard similarity formula and improves it to make the results more accurate. And the carrot2 clustering engine was used to cluster the outcome of separating words in order to increase the clustering speed, and finally worked out the hot topics circulated in current cell phone SMS by the strategy of popularity marking while collecting a certain amount of SMS sample to check the effects in order to conclude the most popular SMS topic. The emulation result declares that the strategy of popularity marking deserves a considerable accuracy. Based on the work mentioned above, this paper has finally designed the SMS Public Information Disposing Platform which is running stabilizing within the test circumstances in the lab.