A Novel Spam Email Detection System Based on Negative Selection

Research output: A Conference proceeding or a Chapter in BookConference contributionpeer-review

21 Citations (Scopus)
311 Downloads (Pure)

Abstract

Nowadays, detecting and filtering are still the most feasible ways of fighting spam emails . There are many reasonably successful spam email filters in operation. However, proactively catching new strains of spam emails, where no previous knowledge is available, is still a major challenge. Negative selection is a branch of artificial immune systems. It has a strong temporal nature and is especially suitable for discovering unknown temporal patterns. This nature makes it a good candidate in quickly discovering and detecting new strains of spam emails. In this paper, we study the feasibility of negative selection in detecting spam emails without using any prior knowledge of any spam emails. We use TREC07 corpus for our experiments. The outcomes, under the assumption of no prior knowledge about spam emails, are very encouraging. We also discuss our findings and point out possible future directions.
Original languageEnglish
Title of host publicationFourth International Conference on Computer Sciences and Convergence Information Technology
Editors Sungwon, Sohn, Kae Dal Kwack, Kyhyun Um, Gye Young Lee, Franz Ko
Place of PublicationSeoul, Korea
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages987-992
Number of pages6
Volume1
ISBN (Print)9780769538969
DOIs
Publication statusPublished - 2009
EventFourth International Conference on Computer Sciences and Convergence Information Technology - Seoul, Korea, Democratic People's Republic of
Duration: 24 Nov 200926 Nov 2009

Conference

ConferenceFourth International Conference on Computer Sciences and Convergence Information Technology
Country/TerritoryKorea, Democratic People's Republic of
CitySeoul
Period24/11/0926/11/09

Fingerprint

Dive into the research topics of 'A Novel Spam Email Detection System Based on Negative Selection'. Together they form a unique fingerprint.

Cite this