Application Of Intelligent Information Processing Technology In The Analysis Of Internet Public Opinions

Internet has continuously becoming the main place where public opinions are produced and spread. In order to enhance the monitoring and control of Internet, the analysis of Internet public opinions is a challenging problem faced by governments. The Public Opinions Monitoring and Analyzing System (POMAS) aims to automatically monitor and analyze the huge amount of Internet pubic opinions in real time. An important function of POMAS is the hot topic detection and analysis of Internet public opinions, which can give the government users a quick understanding and mastery of current hot topics on the Internet. Therefore we in this paper further promote the current researches on the hot topic detection and in-depth analysis of Internet pubic opinions, which are as follows:Firstly, aiming at the practical application of the automatic hot topic detection, we in this paper proposed a practical algorithm to automatically detect hot topics of Internet public opinions. Considering the hot topics' characteristics and people's recognition law, this algorithm introduces several steps to improve the detection effect including: topic sorting, topic combination and adjustment, document elimination and topic description. The experiment shows that the proposed algorithm, in the practical environment, greatly improves the effect and the practicability of hot topic detection.Secondly, for multi-document keyword extraction, we proposed the POS-tagging based phrase method, which effectively utilizes the important global information of many documents, and filters the redundancy information to the best of our ability. The experiments on the Chinese test data demonstrate the validness of our proposed method.Thirdly, for the polarity analysis of the Internet public opinions, in this paper we do the three kinds of tasks: word polarity analysis, document polarity analysis and opinion mining. For the word polarity, we proposed the search engine snippet based method and the Tongyici Cilin based method. For the document polarity analysis, we analyze the polarity word dictionary based method and machine learning based method. For opinion mining, we proposed in this paper the POS-tagging and polarity word dictionary based method.Last, a real application - Public Opinions Monitoring and Analyzing System (POMAS) is introduced and the practical significance of the hot topic detection and in-depth analysis of Internet pubic opinions is demonstrated.