Training the filter system
For users of our Email Security service the system is self learning and there is nothing to do.
If you have a Zimbra mailbox, you can further train the spam filtering system by moving emails into or out of your Junk folder. This works regardless of how you are connecting to your Zimbra mailbox e.g. Webmail, Outlook, ActiveSync etc.
How to fix:
If the email is in your Inbox and it’s not marked as spam and you want the system to learn that it’s actually a spam message, then:
- highlight the email in your inbox
- click on the 'Spam' or 'Junk' button (if you’re using Zimbra webmail)
This will move the spam email from your inbox into your Junk folder.
The other option is to drag and drop the spam email from your inbox to the Junk folder.
If the email is in your Junk folder and it’s been tagged as spam (but is in fact legitimate), and you want the system to learn that it’s not a spam message:
- highlight the email in your 'Junk' folder
- click on the 'Not Spam' button (if you’re using Zimbra webmail)
This will move the legitimate email from your Junk folder back into your inbox.
The other option is to drag and drop the legitimate email from your Junk folder back into your inbox.
What happens when you move an email into or out of your Junk folder?
The process of moving an email into our out of your Junk folder queues the email for learning. This happens immediately so there’s no need to wait before performing other actions on that email – the queue takes time to process but that happens in the background and is unaffected by further actions on the original email.
Moving an email out of Junk causes Zimbra to learn that email as legitimate (ham). Moving an email into Junk accidentally and then out again causes it to be learnt as both spam and ham which effectively cancels each other out.
It can take several emails being learnt from before similar emails are detected as spam/ham – a single reported email simply increases the chance that a similar future email will be detected, it doesn’t cause detection by itself.
This is the case for 99% of the spam filtering rules – multiple rule hits are required before an email is detected as spam. Details are shown in the headers for the rules triggered and the scores assigned.