The Bottom Line
- Spam Monitor uses a sophisticated set of tools to detect spam
- Working with just about any email program, Spam Monitor can configure a few for you
- Spam Monitor queries a sensitive selection of public black lists and can auto-whitelist good senders
- Spam Monitor finds way too much spam that is no spam with its default configuration
- The default data used by the Bayesian filter did not work for me, training it is cumbersome
- Spam Monitor only really works with POP accounts
Description
- Spam Monitor filters spam on multiple POP and IMAP accounts.
- Working as a proxy, Spam Monitor is compatible with any email client.
- Spam Monitor checks any number of public black lists for known sources of spam.
- Additionally, Spam Monitor employs Bayesian statistics to calculate the spam probability.
- Spam Monitor can tag spam in the Subject and in a message header for further filtering.
- Local black and white lists can also be used, and Spam Monitor can auto-whitelist good senders.
- The Bayesian filter can be trained with mail in Outlook Express or Eudora, or with .eml files.
- A plug-in for Outlook makes it even easier to train Spam Monitor and to black- or whitelist.
- Spam Monitor supports Windows 9x/ME/NT/2000/3/XP.
Guide Review - Spam Monitor 2.5 - Spam Filter
At least in my tests, the Bayesian filter found a lot of good mail to be spam using its default set of data. Now, given the nature of Bayesian filters, which are best if they know the patterns of both the good mail and the spam received by each individual user, this is not particularly surprising or disturbing.
Unfortunately training Spam Monitor with your own emails and its mistakes is a bit cumbersome and time consuming. The notable exception is Outlook, where a special Spam Monitor plug-in lets you do the training from a toolbar.
One sensible setting makes the filter learn from mail sent by people you have whitelisted. Spam Monitor can fill up this white list automatically. Of course, Spam Monitor also knows black lists, and it checks a sensible selection of DNS blacklists for known spammers.
For now, it might be a good idea to disable the Bayesian filter and rely on the black lists in Spam Monitor, which can produce not stellar but still solid spam reduction.
