unix电子邮件管理.ppt

* * * * * * * * * * * * * * * * We can find out that the main SPAM mail sources come from e-paper, web-crawling/usenet posing, or dictionary attack and so on. So if you have ever left your email in your personal homepage, or posted messages in the newsgroup, you will probably get the SPAM mails. To construct the email attack ontology, we need to identify the main SPAM mails source first, and then we can follow the METHONTOLOGY approach to construct the ontology step by step. The first step is the “specification” process. Here we need to set the goal of the email attack ontology, and that can tell the other people what the ontology will be used for. Besides the goal, we need to setup the scope of the ontology. It is impossible to build up a complete ontology once, the refinement or modification process is required during the process of ontology construction. With the help of ontology scope, we can know the original goal of the ontology, and that can simplify our refinement. Furthermore, the ontology can be reused, so we always need to use ontology, which is built by others. If someone wants to use the ontology, he can determine whether it fits or not from the scope description. The source of the knowledge indicates where we extract the knowledge from. In the conceptualization step, we should conceptualize the terms and the relationship from the input cases. For example, each email includes “From” , “To”, and “Relay” etc information. We can conceptualize a email as an object and the object contains “From”, “To”, and “Relay” etc. attributes. * * * * * * In “Formalization” step, we need to abstract our class design and build the class hierarchy. The abstraction process would make the classes become high cohesion and low coupling. For example, we can put the general classes in the upper level, and then we can inherit these classes and modify according to our needs to generate specific classes. In implementation step, we make use of Protégé-2000, which is d