Email Thread Identification and Management

Abstract

Nowadays, Email communication is use as primary communication tool in the business domain as well as in education sector. Due to massive incoming emails, overflowing inbox is one of the problems faced by email users. There are several reasons for such a situation, one of them being the unnecessary mass of thread emails. They are retained in inbox even when they are not necessary. Even if this email is deleted from inbox, the next message as thread email will hit your inbox. Wrong use of „reply-all‟ tab adds to this situation called “Email storm”. Thread emails are often generated because of users‟ careless habit to click on „Replyall‟ button. It is almost like a reflex action on their part. This work intends to solve the problem of email storm on two fronts :  Identification of thread emails  Automatically controlling thread email The three datasets Din, Dadm and Dexam from academic domain are used as training data. The experimental outcome shows that „In-Reply-To‟, „References‟ and additionally „threadindex‟ are the dominant features in identifying thread emails. We have used these features to derive thread classification strategy. The developed algorithm is used to test four datasets Dcor, DCS, DF1 and DF2. Using this method accuracy upto 99.91% is achieved. Further, the paper also suggests access control rights strategy to control email storm. The model is proposed for controlling thread emails in education domain. The control mechanism will help system administrators to control email traffic.