How Spam Filters Work
Posted by Jonathan Coupal on: 2006-06-22 18:15:30
Self SEO > Anti Spam Articles
Everyday, e-mail users find their inbox overflowing with messages from people they don't know offering things they don't need. Due to spam, e-mail users waste time everyday deleting junk mail from their inboxes. Sometimes, important e-mails get lost because the capacity of the e-mail account has reached its maximum due to the unsolicited messages.
Even if e-mail users ask the senders of unsolicited messages to stop bothering them, some spam just won't go away voluntarily. The good news is, you can fight spam. There are several techniques available to defend your inbox from unsolicited e-mail including blocking addresses and tracing key words that are generally included in unwanted messages. There are techniques that work automatically and techniques where the user has to train the filter. Listed below are common ways to filter spam and keep it away from your inbox.
White-list and Blacklist:
In this system, also known as blocking, the user organizes a list of trustworthy addresses or domain names and these white-listed e-mails go straight to the user's inbox. On the other hand, the user can blacklist addresses or domain names that send unsolicited messages and make these e-mails will be blocked and go directly to the trash folder. These lists can be maintained at the mail server or on the user's computer.
With this technique an algorithm assigns all characters in an e-mail a numeric value, which it uses to calculate a numeric representation. This "fingerprint" is checked against the database of known spam fingerprints. The algorithm also accounts for whether an e-mail is identical to others received multiple times, generally a good indication of spam.
This kind of filter checks e-mails against a list of spam like keywords and phrases. The more words or phrases are found, the higher the e-mail score.
This filter is trained by the user, who categorizes received e mails as spam or not spam. The filter assigns probability values to each "token" (a word, a phrase, a symbol, or HTML code) based on how often it occurs in spam as opposed to regular messages. An e-mail's score is an average of the token scores. This mechanism has a high rate of success as a filtering technique.
Currently, you cannot completely eliminate receiving spam, but you can utilize spam filters to reduce the amount of messages you get every day. The #1 way to fight spam is to enable a mail filter which will watch your incoming mail, search it for indications of unsolicited content, and help you keep you inbox clean.
ITX offers a robust spam filter that will enable you to manage e-mail more efficiently and effectively. Unlike traditional spam filters, ITX's filter utilizes a combination of mechanisms: advanced statistical analysis based on Bayesian filtering, blacklisting, and anti-virus scanning.