Spam filters, the focus of this guide, are software programs used to detect unsolicited and unwanted email and preventing these messages from getting to a user's inbox.
Most email spam messages are commercial, and most recipients find them to be annoying, intrusive, and potentially dangerous. Many people find that as much as ninety percent, or more, of the email they receive, is spam. It can be difficult to pick the few wanted messages from the many unwanted ones, meaning that important emails are sometimes missed or deleted as spam. Spam can also be dangerous because they may contain links that lead to phishing sites or sites that host malware, or they may include malware or viruses as attachments.
Spammers collect email addresses from several places, including social media, websites, customer listings, and viruses that harvest user address books. Collected email addresses are often sold to other spammers.
A spam filter is a software program designed to detect unsolicited and unwanted email and to prevent these spam messages from getting to the user's inbox. Like other filtering programs, a spam filter uses certain criteria upon which to base its judgments.
Simple filters might look for specific words or phrases in the subject line of email messages, then excluding these from the user's inbox. However, spammers quickly learn to use different words and phrases, while legitimate messages may be blocked. Other types of spam filters, such as the Bayesian filters, Markovian discrimination, or other heuristic filters, differentiate spam from legitimate messages by suspicious word patterns or word frequency.
Other methods include whitelisting, in which identified sender IP addresses, email addresses, or domain names are passed through. Blacklisting may also be used. Also known as blocklisting, this is a method that specifically excludes certain email addresses, IP addresses, or domain names. In most cases, blacklisting makes use of public anti-spam lists that are built through the experiences of other users.
Many, and perhaps most spam filters employ multiple methods, some of them optional. Whatever the method clients designed to filter spam from email are the focus of topics in this category.
Recommended Resources
Developed by Apache Software Foundation, SpamAssassin is a software program designed for email spam filtering, accomplishing it through a variety of spam-detection techniques, such as DNS-based, fuzzy checksum-based, Bayesian filtering, external programs, blacklists, and online databases. Available under the Apache License, it can be integrated with the mail server to automatically filter mail for a site. Its features are listed, along with development notes and sponsorship credits.
https://spamassassin.apache.org/
Written in Perl, MailScanner is an open-source (GPL) email security system for use on Unix gateways. MailScanner GUIs include eFa Project (CentOS), Mailborder (Ubuntu), MailWatch (Red Hat, Debian, Ubuntu, SuSe, FreeBSD), and Baruwa (Debian, Ubuntu, CentOS). Requirements, installation instructions, and post-installation instructions are provided, along with a download page, updates, documentation, and support options. A manual can be downloaded.
https://www.mailscanner.info/
Written in Perl, POPFile is a free and open-source, cross-platform mail filter that uses a naive Bayes classifier. Typically, it is used to filter spam emails, but it can be used to sort mail into defined buckets or categories. Available under the terms of the GNU General Public License, it may be freely downloaded from the site, along with the source code. A user manual, FAQ, and development notes are posted to the site. An online discussion forum is available for support.
http://getpopfile.org/
Created in 2002, SpamBayes is a Bayesian spam filtering program, differing from the traditional Bayesian model is in the emphasis on testing newer approaches to scoring messages. SpamBayes consists of a message classifier, its core code, as well as a number of other applications, mostly operating on the client-side. Compatible with Mac, Windows, and Unix/Linux-type operating systems, it may be downloaded from the site. User documentation and reviews are included.
http://spambayes.sourceforge.net/
Developed and distributed by Axaware, SpamBully is an anti-spam software that uses Bayesian filtering to separate legitimate from illegitimate emails, using allow/block lists to determine which emails should be sent on to the user’s inbox. Compatible with the Windows operating system, the software works with Outlook, Live Mail, Outlook Express, Windows Mail, and IMAP. Purchases may be made online by credit card or Paypal. A demo and help files are available.
https://spambully.com/
Spamhilator uses a variety of filters to achieve its anti-spam functions. Its learning filter uses the Bayesian filter, calculating a spam probability for each email, as well as a spam word filter that searches messages for known keywords, allowing for user-defined words and regular expressions, and a link filter and rule filter, each of which can be configured by the user. Available free, in English and German, the program works in nearly every email client. Plugins and add-ons are available.
https://www.spamihilator.com/
Rather than using pattern matching and a set of human-generated rules, SpamProbe relies on a Bayesian analysis of the frequency of words used in spam and non-spam emails received by an individual. Its origins are highlighted, development notes and a list of features is posted to the site, along with the known platforms on which it will compile. The software client is free to download from the site, and its source code is distributed under the terms of the Q Public License.
http://spamprobe.sourceforge.net/
