The Design And Implement Of Public Opinion Information Process System For Bulletin Board System (BBS)

With the rapid development and expansion of the Internet in the global， networkmedia has become the “fourth media” only after newspaper，television and radio. It is aplatform for dissemination of public opinion that can not be ignored. The system ofpublic opinion in network (network opinion system) is a kind of virtual cyberspace inthe Internet， which takes netizens or netizens groups as node and informationdissemination path as tie. It is complicated， self-organized， social， etc. Publicopinion in network (network opinion) has the following features： the openness， theinteractivity， the secretiveness， the complexity， etc. It is a double-edged sword. Onone hand， it can help people communicate， convey their emotions， have positiveimpact on making decisions and administrations of social management department，buton the other hand， some bad or illegal news also retail through network and havenegative impact on society. So， using the law and social morality to regulate publicopinion in network， guiding it with method of combining the technology andmanagement， is significant to promote innocent Internet culture and reduce oreliminate the negative impact which bad Internet culture brings to the social audience，and also， an important practical problem for governments and society which arouseswide concern. We need to study observation and analysis method of evolution andstate-change of public opinion in network， and find out the object and timing ofguiding network opinion. The study of guiding mechanism and method has become animportant burning problem.In this thesis， we design a specified information process system to the publicopinion via Internet. As a base of the whole study， we deeply research the methods ofcollecting data in Internet. And then， we widely study the various methods of textanalysis， for example， full text indexing， automatic digesting， text classifying，text assembling， which give our system a solid basis， specially the governance ofpublic opinion via Internet.Firstly， we explore the key techniques for the requirement of design system， forexample， the collecting data of BBS， the smart analyzing data of BBS. Especially， the smart analyzing data of BBS includes various key techniques. There are four coretechniques for the system， which are full text indexing， automatic digesting， textclassifying， text assembling. Secondly， based on the above techniques， the systemis designed. Furthermore， we implement the system depend on concreting the abovekey techniques， for example， page-based collecting data of BBS， keywords andcontext-based automatic digesting， Bayes-based content classifying， andK-Means-based information assembling.Finally， the described methods are adopted to test our system. The results showthat the theory and method of network opinion state observation and guidance hastheoretical and engineering directive significance.