The Internet unhealthy information is characteristic of extensiveness，fast speed，multi channels，secrecy and social hazard. It ranges from its original simple text to diverse forms，including words，pictures，audio，video，etc. There are different ways to identify and supervise different forms of information.
　　2.3.1Text information：Currently，four methods are mainly used：key word matching method，latent semantic indexing method，neural network method and identification method similar to information retrieval. The keyword matching method，a quite simple one，is usually used to build Boolean logic models and the vector space model. However，owing to the lack of contexts，it is hard to identify information accurately. The latent semantic indexing method is involved in the processing of natural language through establishing a document space based on latent semantic indexing. Documents of similar themes are near to each other in the space. It identifies information according to the threshold that has been set.This method is rather complicated. The system performs slowly and its capability to identify will be decreased with the addition of new words. The neural network method is to imitate the structure of human brain and the function of the nerve system. It virtualizes the information that the users are interested in to be an artificial neural network，and combines with various mathematical algorithms to realize information identification. This possesses a certain fault tolerance and flexibility and can update automatically，but it is more complex and slow.
　　2.3.2 Multimedia information：In addition to text information，multimedia information also includes image information，audio information and video information，which is more difficult to identify and supervise than text information. Multimedia information is a time series of stream，so it can not deal directly with a long video or audio. Thus，multimedia data should be divided into different lengths and be classified with the multimedia physical units. Identification and supervision of the images are mainly two kinds：text-based and content-based. For the identification and supervision of the text-based multi-media information，it is to build an index database similar to a text document information retrieval system，marking with keywords，identify through text matching. This method is simple and inexpensive，but it can only be semi-automatic，remaining to improve its accuracy. Identification of content-based multimedia information first extracts the features of the multi-media (to represent the original signal data form，such as the extracting of color，texture and movement from the image and video; sound，pitch and tone from the audio)，and then system recognizes the semantic categories and sets up index for them. This demands higher system storage and counting ability. 　At present，the text testing technology is more advanced，and can achieve automatic testing with high precision and efficiency as well. As for images，audio and video testing technology，the input and the outcome are not proportional. It often combines management with technology.
　　2.4 The transmitting ways of unhealthy information
　　As the business of the Internet continues to expand，the transmission of information has developed from the simplest HTTP web browser to P2P，IM etc. Difficult measures of identification and supervision should be adopted to counter with unhealthy information transmitted by different ways.
　　2.4.1 HTTP method：This is the most popular way that transmits unhealthy Internet information. Its identification and test are fulfilled by active pull testing and passive supervising，usually with the help of Web Worms together with DPI (Deep Package Instruction) to monitor the flow and content of the HTTP.
　　2.4.2 P2P method：This is called “point to point” method which is commonly applied in data downloads，video streaming etc. Users download，at the same time，computers upload. In this way，data is scattered in a number of nodes so that it is difficult to test and filter. Supervision is the main measure to manage this type.
　　2.4.3 IM method (Instant Message)：It is a convenient way of communication for the users to communicate online，transfer files，voice chat and video conference. Unhealthy information is commonly transmitted in this way，and Internet service providers provide only the bottom of the network，it is difficult to identify the unhealthy information through the supervision of the flow. So it usually reviews content from ICP side.
　　3.The countermeasures dealing with the proliferation of Internet unhealthy information
　　3.1 Establish sound laws and regulations concerning information security and build an efficient management mechanism
　　The Internet unhealthy information is characteristic of extensiveness，fast speed，multi channels，secrecy and social hazard. It becomes more difficult to supervise the Internet unhealthy information because of differences in cultural backgrounds，legal environment，social trends and political systems. At present，the measures to deal with unhealthy information in China are still artificial and passive ones and it is dealt with only if it has caused damage. How to prevent and nip unhealthy information in the bud is the key to Internet governance.(责任编辑：南粤论文中心)转贴于南粤论文中心: http://www.nylw.net(代写代发论文_毕业论文带写_广州职称论文代发_广州论文网）