Advertising Email Management Using Multi-Level Reputation System

Currently, a large number of emails from sources either wanted or unwanted sites are sent to our mail box every day. Most of emails from unwanted sites are left unread and finally deleted without opening. Emails even from some subscribed sites that are no longer interested are also left unread. The problem is that it could not have an easy way to unsubscribe or make unwanted senders stop sending emails to us. Thus, we propose Email Management Framework to help recipients gradually reducing emails from unwanted senders automatically. This framework has used centralized user feedback databases and multi-level reputation system. We divided false reporting feedback into positive and negative feedback to prevent sending mistaken emails to recipients. From our experiments, the results show that the proposed framework can increase accuracy rate of classification 6.20% compared to traditional process. Furthermore, the framework can provide the way to manage email sending automatically by using feedback from recipient's behavior. Recipients are ensured that they would finally receive only wanted emails. Unwanted emails will gradually be reduced automatically. Additionally, for sender's perspective, the framework also provide value-added information who are their real customers and are still interested in their information.