(
It would be great if SpamBayes' classifying power could be used to further categorize and prioritize good messages, and plug-ins for programs other than Outlook and Mozilla Thunderbird would be welcome.
- SpamBayes detects spam precisely and efficiently
- A POP proxy and IMAP module make SpamBayes work with most email accounts
- Easy to use plug-ins for Outlook and Mozilla Thunderbird are available
- SpamBayes cannot further classify or prioritize good mail
- With email programs into which SpamBayes does not plug directly, setup can be a bit tricky
- SpamBayes filter spam in POP and IMAP email accounts, and plugs directly into Outlook and Mozilla Thunderbird.
- SpamBayes uses Bayesian statistics to detect spam, learning with every mail (and as you correct it).
- A POP proxy adds a classification header the email program can use to move incoming junk using a rule.
- To train SpamBayes, you can use a web interface or forward mis-classified to a special SpamBayes address.
- With IMAP accounts, SpamBayes moves spam right at the server. To train the filter, you move the messages, too.
- In Outlook and Mozilla Thunderbird, you can use a toolbar to train SpamBayes.
- Outlook can be configured to show the SpamBayes spam probability in the mailbox summary.
- SpamBayes can display a detailed analysis for each email.
- SpamBayes supports Windows 98/ME/NT/2000/3/XP/Vista.
After a just bit of training, SpamBayes detects junk mail reliably. With IMAP accounts, it moves the junk to a special folder automatically. Acting as a POP proxy in between your email program and the mail server, SpamBayes adds a marker that lets the email program filter the junk further.
To correct error, you can use SpamBayes's web interface or forward the messages to a training address with POP accounts and just move to distinct folders with IMAP. Outlook and Mozilla Thunderbird users have it particularly comfortable, too: SpamBayes plugs into these email programs for easy setup and training.
The option to put emails whose spam probability is somewhere between good mail and junk in a special folder make it easy to continue training and correcting (though that will not be necessary often) SpamBayes' decisions. Forget about the other, non-incremental training mode, though.
If you're curious what spam is made of, SpamBayes can show you a detailed analysis of a message's overall spam probably and its individual word counts.
It would be great of SpamBayes wonderful email classifying talent could be harvested for prioritizing and categorizing good mail further. Direct support for more email programs would be nice, too.
