Augmented Interval Type-2 Fuzzy Set Methodologies for Email Granulation

Abstract

Email, as one the most popular Internet service is confronted with the plague of spam, which results in bandwidth, time and money waste. Moreover spam has evolved into a true security issue that enforces organizations to fight back in order to address security measures confidentiality, integrity and availability. To this end, we have proposed a dynamic model to classify incoming messages into five granules namely, spam, suspicious-spam, suspicious, suspicious-non-spam and non-spam, using interval type-2 fuzzy set methodologies augmented with the concept of general intervals. Despite the intrinsic complexities of higher order fuzzy sets, the error ratio of misclassification of the proposed method is noticeable. However it should be stressed that no single method can achieve one hundred percent precision, the proposed model should be used in conjunction with other complementing technologies.