Common Anti-Spam Filtering Techniques

Email spam has been a problem for almost twenty years and continues to plague us today. This problem would be far worse if we didn’t have anti-spam software. Two common filtering techniques for blocking email spam will be discussed below:

List Filters

The simplest anti-spam technique makes use of lists to either allow email into or block email from an inbox. A list of domains and email addresses that are known to be used for spam is called a blacklist. Additional email addresses can typically be added to the list by flagging a given email as spam. Blacklists can be defeated through the use of multiple email addresses. Addresses can also be spoofed by the spammer to get past blacklists.

The white-list is the opposite of a blacklist in that it contains non-spam email addresses. Emails with addresses on this list are always allowed through. Sometimes there is an option to only allow white-listed email. This is often impractical because it doesn’t allow for legitimate addresses that aren’t on the list. Spammers can defeat white-lists by spoofing email addresses harvested from address books.

Another type of list contains words and phrases commonly used by spammers. The subject line and content of email are checked against this list and blocked if the spam words or phrases are found. This type of filter can be circumvented by altering phrases and the spelling of words. False positives can occur when an email is discussing a spam topic for legitimate reasons.

Bayesian Filters

The Bayesian filter uses a statistical technique that compares the content of an email message to characteristics commonly found in spam email. It then assigns a probability of the email being spam which depends on how closely the content matches the characteristics of a spam message. The threshold probability that determines whether a message is spam, can be changed depending on how aggressively one wants the filter to flag spam. The setting is a trade-off between reducing spam that gets through the filter and the number of false positives.

When the user flags email as spam, its characteristics are added to the Bayesian filter’s database. In this way, the filter continually adapts to the changing nature of spam.

For more information on how our services can help your business, please contact us.