The Internet has become an important part of human society with its popularisation and application in various fields of our society. Web pages which spread in the Internet are in a variety of forms. It is an important topic in Internet research to make effective distribution regulation on the web content, especially on the harmful content. The paper analyzes the various measures which are taken by the countries all over the world to filter the harmful information in Internet and then proposes an information filtering system which combines the Web Rating System and SVM classification algorithm. This information filtering system can rate the web pages automatically according to the classification templates which have been trained by the rating system and finally filters the harmful information. In the implementation process, the paper also introduces some practical security measures which can ensure the whole system to work reliably.We mainly made the following contributions in this paper:(1) We analyze the current filtering technologies and the obstacles in the web content rating system and put forward an information filtering system which combines the Web Rating System and web pages automatic classification.(2) We implement the program which can automatically classify the unrated web pages. The program mainly includes the follow parts:the web information transformation, characteristic vector calculation, classification training, web page classification and rating label generating. We also collect a large number of web pages to evaluate the filtering system.(3) We adopt the digital signature to ensure the reliability of the rating label and use the digest of information to ensure the consistency of rating label with the corresponding web page. We describe the method and the steps in details. |