Ontology based classification and categorization of email

Due to the growing popularity of email, spam email continues to evolve and grow. All the unsolicited email and the emails not intended for a particular user are classified as Spam Email. The spammers are very intelligent in concealing their mail from spam filters. The spam creates increase in the network traffic and unnecessary wastage of time for the user. One of the foremost challenges in designing a spam filter is that the spam is user dependent. A mail that appears to be legitimate to one user may appear spam to another and vice versa. This imposes a requirement on the design of spam filter that it should be made user defined and user controllable. It is possible to locate all the mails related to a particular category if the number of emails in the inbox is less. As the number of emails in the inbox increases it becomes time consuming and difficult for a user to search all the emails belonging to a particular category. A filter similar to spam filter can be designed to make such categorization automatic. The paper proposes a method to create an email classification filter to satisfy both the above objectives. It uses ontology for understanding the content of the email and Bayesian approach for making the classification.