A spam filter is a software program that sorts incoming mail in order to identify and pull out junk mail, also known as spam. It can be installed on an Internet mail server, on a private network server, or on a personal computer. Spam is not only bothersome, but it can be used for spreading malicious code like viruses and Trojans, and for perpetuating phishing scams. For these reasons and more, a spam filter is a great way to help protect a computer or network and cut out junk mail.
In order to detect spam, the filter compares parameters in incoming mail to lists of configurable rules. For example, it can be set to check the subject heading of incoming mail for terms associated with pornography, pharmaceuticals and other common spam products. The sender field might also be filtered for ranges of IP address associated with spammers or marketers.
Over the years, spammers have become more sophisticated in their attempts to get junk mail past spam filters. Spam will often include innocuous or even personable subject headings such as "Hi!" or "Regarding your inquiry," making it difficult for a general ISP filter to tell spam from legitimate email. Therefore, some spam still gets passed from Internet mail servers to connected private networks and to end-users.
For this reason, many network administrators install a spam filter on the office or company network, as do end-users at home. In these cases, the filter can be highly configured to catch junk mail that makes it past the more generally-configured programs located online. Like other software programs, spam filters are not all equal. Currently, the most popular is an "intelligent" program based on Bayesian statistical methodology. It compares incoming mail to a set of user-defined rules, then identifies spam based on mathematical probabilities.
In the simplest configuration, a spam filter passes junk mail to a spam folder, and legitimate mail to the inbox. The user can review the assignments and correct mistakes. The program will remember corrections and avoid making those mistakes in the future. In other words, a Bayesian program is "taught" what to look for, and "learns" from its mistakes. Before long, the filter will make few mistakes and can operate virtually user-free. One popular free spam filter boasts a near-perfect record after an initial "training" period.
A good filter will also use a "white list" to provide protection from phishing schemes. People who do online banking, for example, can enter the bank's correct website address in white list. The spam filter will check incoming mail for misspelled or similarly spelled addresses and reject them as phishing attempts.
There are several spam filtering programs available online, most with a 30-day free trial period.