- \n
- Whitelist management<\/li>\n
- Hash-sharing systems (DCC<\/a> & Razor<\/a>)<\/li>\n
- Token-based regex\u00a0matching<\/li>\n
- Markup filtering<\/li>\n
- Bayesian<\/a> filtering<\/strong><\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
All steps in\u00a0the filtering process are automated, except for Bayesian filtering<\/strong> that works by both\u00a0automatic learning and manual learning. This covers how\u00a0to train your filter to\u00a0improve\u00a0filter performance.<\/p>\n
How it works<\/h2>\n
Bayesian filtering breaks an e-mail down into individual words, then compares the probability of words in legitimate e-mail and spam. If certain words or phrases such as “Dr. Oz”, “Solar Panels”, and “Viagra” appear more frequently in e-mail identified as spam, then that e-mail that contains such phrases is likely to be spam as well. Likewise phrases that contain, “Monday”, “Synergism”, and “Ocelot” may be less likely to contain spam based on training data. E-mails that come in with those words are rated more favorably as non-spam and, therefore, less likely to be delivered to your Spam folder<\/a>.<\/p>\n
How\u00a0to use it<\/h2>\n
<\/p>\n
Training by IMAP folder<\/h3>\n
![\"Drag](\"https:\/\/kb.okra.host\/wp-content\/uploads\/2014\/11\/autospam-learning-folder.gif\")
<\/a><\/p>\n