Abstract

The Internet is popular, with email use functioning as the major Internet activity. However, spam has recently become a major problem impeding the use of email. Many spam filtering techniques have been implemented so far. Most current anti-spamming techniques filter out junk emails based on email subjects and body messages. Nevertheless, subjects and email contents are not the only cues for judging spam. This investigation presents a statistical analysis of the header session messages of junk and normal emails, and explores the possibility of utilizing these messages to perform spam filtering. The message head session, including the sender's mail address, receiver's mail address and time, which is of little interest to most users, also provides further information for anti-spamming purpose. A statistical analysis is undertaken on the content of 10,024 junk emails collected from a Spam Archive database, and 599 regular emails in company with 635 solicited listserv or commercial emails contributed by volunteers. Content analysis results demonstrate that up to 92.5% of junk emails are filtered out when utilizing the message-ID, mail user agent, and sender and receiver addresses in the header session as cues. Additionally, the proposed approach may induce a low block error rate for normal emails for the sample utilized in this investigation. This low rate of over-block errors is a significant merit of the proposed anti-spamming approach. The proposed approach of utilizing header session messages to filter out junk emails may coexist with other anti-spamming approaches. Therefore, no conflict arises between the proposed approach and existing spam prevention approaches.

Keywords: Spam; Unsolicited email; Junk mail; Email address; Filter

Article Outline

1. Introduction

2. Anti-spamming approach

2.1. Bulk email filtering

2.2. Filtering by keyword

2.3. Blacklist

2.4. Whitelist

2.5. Sender address validity

2.6. Receiver address as cue

2.7. Mail user agent as cue

2.8. Message-ID as cue

3. Efficiency of anti-spam techniques

4. Content analysis for junk and normal emails

4.1. Sender addresses

4.2. Receiver addresses

4.3. Mail user agent

4.4. Message-ID

5. Using header session message to block spam

6. Discussion

Acknowledgements

References

Vitae