Partner Perspectives  Connecting marketers to our tech communities.
SPONSORED BY
3/14/2017
11:00 AM
Pieter Arntz
Pieter Arntz
Partner Perspectives
Connect Directly
Twitter
RSS
50%
50%

7 Things You Need to Know about Bayesian Spam Filtering

Knowing how spam filters work can clarify how some messages get through, and how your own emails can avoid being caught.

Bayesian spam filtering is based on Bayes rule, a statistical theorem that gives you the probability of an event. Bayesian filtering is used to give you the probability that a certain email is spam.

1. The Name
It’s named after the statistician the Rev. Thomas Bayes, who provided an equation that allows new information to update the outcome of a probability calculation. The rule is also called the Bayes-Price rule after the mathematician Richard Price, who recognized the importance of the theorem, made some corrections to Bayes’ work, and put the rule to use.

2. Spam
When dealing with spam the theorem is used to calculate a probability about whether a certain message is spam. The probability is based on words in the title and message, derived from messages that were identified as spam and messages that were identified as not being spam (sometimes called ham).

3. False positives
The objective of the learning ability is to reduce the number of false positives. As annoying as it might be to receive a spam message, it is worse to not receive a message from a customer just because he used a word that triggered the filter.

4. Scoring
Other methods often use simple scoring filters. If a message contains specific words a few points are added to that messages’ score and when it exceeds a  certain score, the message is regarded as spam. Not only is this a very arbitrary method, it’s also a given that it will result in spammers changing their wording. Take for example “Viagra” which is a word that will surely give you a high score. As soon as spammers found that out they switched to variations like “V!agra” and so on. This is a  cat and mouse game that will keep you busy creating new rules.

5. Learning
If the filtering is allowed for individual input the precision can be enhanced on a per-user base. Different users may attract specific forms of spam based on their online activities. In other words,  what is spam to one person is a “must-read” newsletter to the next. Every time the user confirms or denies that a message is spam, the filtering process can calculate a more refined probability for the next occasion.

6. Poisoning
A downside of Bayesian filtering, in cases of more-or-less targeted spam, is that spammers will start using words or whole pieces of text that will lower the score. During prolonged use, these words might get associated with spam, which is called poisoning.

7. Bypasses
A few methods to bypass “bad word” filtering.

  • The use of images to replace words that are known to raise the score

  • Deliberate misspelling, as mentioned earlier.
  • Using homograph letters, which are characters from other character-sets that look similar to letters in the messages’ character set. For example, the Omicron from the Greek looks exactly the same as an “O," but has a different character encoding.

Bayesian filtering is a method of spam-filtering that has a learning ability, although limited. Knowing how spam filters work will clarify how some messages get through, and how you can make your own mails less prone to get caught in a spam filter.

Links to more information:

Was a Microsoft MVP in consumer security for 12 years running. Can speak four languages. Smells of rich mahogany and leather-bound books. View Full Bio
Comment  | 
Print  | 
More Insights
Comments
Newest First  |  Oldest First  |  Threaded View
Microsoft, Mastercard Aim to Change Identity Management
Kelly Sheridan, Staff Editor, Dark Reading,  12/3/2018
Windows 10 Security Questions Prove Easy for Attackers to Exploit
Kelly Sheridan, Staff Editor, Dark Reading,  12/5/2018
Starwood Breach Reaction Focuses on 4-Year Dwell
Curtis Franklin Jr., Senior Editor at Dark Reading,  12/5/2018
Register for Dark Reading Newsletters
White Papers
Video
Cartoon Contest
Write a Caption, Win a Starbucks Card! Click Here
Latest Comment: I guess this answers the question: who's watching the watchers?
Current Issue
10 Best Practices That Could Reshape Your IT Security Department
This Dark Reading Tech Digest, explores ten best practices that could reshape IT security departments.
Flash Poll
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
CVE-2018-20000
PUBLISHED: 2018-12-10
Apereo Bedework bw-webdav before 4.0.3 allows XXE attacks, as demonstrated by an invite-reply document that reads a local file, related to webdav/servlet/common/MethodBase.java and webdav/servlet/common/PostRequestPars.java.
CVE-2018-20001
PUBLISHED: 2018-12-10
In Libav 12.3, there is a floating point exception in the range_decode_culshift function (called from range_decode_bits) in libavcodec/apedec.c that will lead to remote denial of service via crafted input.
CVE-2018-20002
PUBLISHED: 2018-12-10
The _bfd_generic_read_minisymbols function in syms.c in the Binary File Descriptor (BFD) library (aka libbfd), as distributed in GNU Binutils 2.31, has a memory leak via a crafted ELF file, leading to a denial of service (memory consumption), as demonstrated by nm.
CVE-2018-19991
PUBLISHED: 2018-12-10
VeryNginx 0.3.3 allows remote attackers to bypass the Web Application Firewall feature because there is no error handler (for get_uri_args or get_post_args) to block the API misuse described in CVE-2018-9230.
CVE-2018-19653
PUBLISHED: 2018-12-09
HashiCorp Consul 0.5.1 through 1.4.0 can use cleartext agent-to-agent RPC communication because the verify_outgoing setting is improperly documented. NOTE: the vendor has provided reconfiguration steps that do not require a software upgrade.