The Network Public Opinion Monitoring System Research And Exploitation

With the development of the internet as a fourth medium following the three major traditional media : newspapers, radio, television, its special effects are becoming increasingly apparent, the network of public opinion has become the birthplace of social public opinions and amplifier. How to correctly guide the network public opinion , to prevent the spread of bad situations which might bring to the ruling party and governments at all levels unprecedented challenges. It is nessecery to effectively monitor network public opinion, and reasonably guide people's attitudes towards hot social issues, and it is quite urgent to improve the monitoring capability of the network monitoring system .Some features of network public opinion have limited the application of traditional techniques(chinese word segmentation,text mining), for example , network public opinion has the character of being hidden somewhere so that it cannot be easily noticed by people and it is also very extensive . Meanwhile these features of network public opinion also put forward higher requirements for network monitoring system of public opinion . This thesis is centred around the desigh of network monitoring system of public opinion and how to put it into practice in order to operate research work with the aim of enhancing traditional text mining applications in the system .This thesis has designed and carried out the following contents :Through analysis of the demand for network public opinion , to build an overall system framework for the network public opinion system , effectively solve problems referring to its structural design and system processes and verify its legitimacy and effectiveness by practice .In connection with related technologies at home and abroad, to optimize the analysis and desigh of module respectively , completing a network monitoring system of public opinion with practical significance. To effectively improve the K-means clustering algorithm and put forward cohesion proposed k-means text clustering algorithm , starting with optimizing the text mining technology . Through the comparison between simulation examples and traditional algorithms , it is verified that the text cohesion k-means clustering algorithm has more effectiveness and superiority in raising the accuracy and efficiency of finding topics .