Content based email classification system by applying conceptual maps

This paper discusses the possibility of adopting the concept of knowledge based systems [KBS], in general, and conceptual maps, in particular, in email classification system. Needless to say that email has the potential to improve efficiency and reduce costs involved in communication. Even after the advent of newer technologies such as instant messaging and VoIP, email remains the top most application of the Internet, intranets and extranets. Though the email system is popular, powerful, cost-effective and efficient, it has some shortcomings viz., [a] it remains an unmanaged medium and [b] it is a vast store of unstructured information. Hence, email database will effectively be used by the knowledge workers only if: [a] it enables classification within stringent legitimate frame work [b] it can be searched/sorted and mined in ways that make it useful. Email classification, a technique for adding metadata and visual labels to email, offers an effective strategy for managing and controlling email. Most of the existing systems are based on term-based classification schemes. To achieve better accuracy, semantic information must be made use of. This proposed system could be an add-on application of existing computational as well as term-based methods.