CRM114 versus Mr. X: CRM114 notes for the TREC 2005 spam track (2005)
| Venue: | In Proc. 14th Text REtrieval Conference (TREC 2005 |
| Citations: | 5 - 0 self |
BibTeX
@INPROCEEDINGS{Assis05crm114versus,
author = {Fidelis Assis and William Yerazunis and Christian Siefkes and Shalendra Chhabra},
title = {CRM114 versus Mr. X: CRM114 notes for the TREC 2005 spam track},
booktitle = {In Proc. 14th Text REtrieval Conference (TREC 2005},
year = {2005}
}
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
Abstract: This paper discusses the design decisions underlying the CRM114 Discriminator software, how it can be configured as a spam filter, and what we may glean from the preliminary TREC 2005 results. Unlike most other filters, CRM114 is not a fixedpurpose antispam filter; rather, it’s a general purpose language meant to expedite the creation of text filters. The pluggable CRM114 architecture allows rapid prototyping and easy support of multiple classifier engines; rather than testing different cutoff parameters, the CRM114 TREC test set tested different classifier algorithms and learning protocols. Introduction and History: The CRM114 Discriminator was initially begun as a rough design in 1998 as a hypothetical filter design, meant to categorize incoming mail into a significant number of different “topics ” (and with spam as an acknowledged target, but by no means the sole target). One of the design assumptions that CRM114 has held from day 0 has been that single word features are not as important as Nfeature tuples. The initial CRM114 test code worked on Nletter tuples, rather than Nword tuples, but the groundwork had been laid. Testing against the very poorly camouflaged spam of
Citations
No citations identified.
