A Collaborative Reputation-Based Vector Space Model for Email Spam Filtering

In this paper, we propose a novel Collaborative Reputation-based Vector Space Model (CRVSM) for detection of spam email. CRVSM uses a vector space model for representing the feature vectors in multidimensional vector space in order to detect the spam emails in large space. We cluster the emails into five clusters so as to reduce the email spam detection time. To reduce the number of false positives and false negatives, we calculate maximum similarity measure with maximum and minimum threshold range. Moreover we use a reputation evaluation function which determines the reporter's trust level in validating an email as spam or non-spam. The CRVSM approach achieves good efficiency while obtaining good reputation result in Email spam detection. The performance of CRVSM model has been evaluated using metrics such as false positive rate, false negative rate, detection accuracy and detection time. The performance results clearly show that CRVSM accurately detects the incoming emails as spam or non-spam with less FPR and FNR values thereby achieving a high efficiency with short detection time and outperforms the existing detection protocols.