A Proposed Model for Malicious Spam Detection in Email Systems of Educational Institutes

The cheapest form of communication in the world today is email, and its simplicity makes it vulnerable to many threats. One of the most important threats to email is spam, unsolicited email, normally with an advertising content sent out as a mass mailing. Malicious spam is spam with malicious content in forms of harmful attachments or links to phishing websites. In the case of educational institutes, malicious spam threatens the privacy and security of large amount of sensitive data relating to staff and students. Hence, a system that can automatically learn how to classify malicious spam in educational institutes is highly desirable. In this paper, we aim to improve detection of malicious spam through feature selection, with focus on the educational field. We propose a model that employs a novel dataset for the process of feature selection, a step for improving classification in later stage. This dataset is unprecedented as no research in the literature was intended to serve malicious spam detection in a specific domain or field such as the educational field. Feature selection is expected to improve training time and accuracy of malicious spam detection.