The bayesian analysis wizard

The bayesian analysis wizard is a tool to update the spam signature set used by the bayesian analysis filter with custom spam and legitimate e-mail. Although GFI MailEssentials ships with a default spam signature database and learns legitimate mail by adding your outbound mail to its signature set, you could use the wizard for the following reasons:

1. Instead of waiting for bayesian filter to learn from outbound mail (takes about 4 days), run the wizard so that the bayesian filter is more effective immediately.

2. If you are not in an english speaking country, you can use the bayesian filter to identify spam mail in your language. You can use the bayesian wizard to add signatures for for example german spam mail.

Installing the bayesian wizard on another machine

On many occasions you probably want to run the bayesian wizard on a separate machine then the MailEssentials server. To do this:

1. Copy the file bayesianwiz.exe, located in the MailEssentials\bsw folder, to the machine on which you wish to install the bayesian wizard.

When running the bayesian wizard on a separate machine, you will have to copy the bayesian wizard output, the so called .bsp file back to the GFI mailessentials machine to the right location, so that it will use your bsp file and not the default one included in the installation.

Preparing spam and ham mail stores

The Bayesian analysis wizard will analyze a source of legitimate mail (for example your Outlook sent items folder) also known as HAM, then analyze a source of spam mail (which for example you stored in outlook or exchange, and subsequently create a new bayesian spam profile (bsp file) which can be used by the GFI MailEssentials 9 filter on the server.

Identify a folder in which you have legitimate mail stored. The easiest choice is a user's sent items folder.

Create a folder in which you will copy spam mail, or alternatively use the GFI created spam profile.

You can access these sources of SPAM and HAM either through your Microsoft Outlook profile, or by storing it on your exchange server and connecting directly to your Exchange Server.

Running the analysis wizard

1. From the GFI MailEssentials program group, run the Bayesian spam wizard. Click Next to continue.

2. Now choose to create a new bayesian spam profile file or update an existing one. The Bayesian filter depends on the Bayesian spam profile (bsp) file to identify spam. This bsp file should be stored on the machine running GFI MailEssentials. Choose whether to:

Create or Update a Bayesian spam profile (bsp) file. After completing the wizard, you will have to ensure that this file will be copied to the MailEssentials\data directory. (GFI MailEssentials will have to be stopped temporarily). Select the file to update or create a new one by clicking on the <...> button.

Update Bayesian spam profile file used by the Bayesian filter directly. This requires that you run the wizard on the machine on which GFI MailEssentials is installed. This option will update the file being used by GFI MailEssentials. If you are running the wizard remotely this option will be greyed out.

Selecting how to access the HAM and SPAM stores

3. Now select how the wizard will access the sources of SPAM and HAM that you have created. The wizard can do this either through your Microsoft Outlook profile, or by connecting directly to your Exchange Server. If you select the latter, you have to store the mail on the Exchange Server and specify logon credentials. Alternatively you can select not to update legitimate mail (ham) in the Bayesian Spam profile. In this case, only the spam profile will be updated in the resulting bsp file (you do not need to update your HAM profile very often).

If you wish to use your Microsoft Outlook profile, ensure Microsoft Outlook is running.

4. After you have selected how the wizard should connect to your mail stores, you have to select the appropriate folders that contain the HAM mail. We recommend using the sent items folder, since these always contain HAM.

5. Now you have to idenfiy SPAM mail. Select how you are going to access the mail store with spam. The process of doing this is identical to identifying HAM mail. You can either select a folder in which you have stored SPAM mail (same process as for legitimate mail) or you can download latest Spam profile from the GFI website. GFI maintains a spam profile file which it regularly updates by collecting mail from leading spam archive sites. Alternatively you can select not to update Spam in the Bayesian Spam profile. (Only ham will be updated in bsp)

6. If you have selected a folder in which you stored SPAM, you have to select the appropriate folders that contain the SPAM mail. Click Next to continue

7. Now the wizard is ready to create/update the Bayesian spam profile file. This process may take several minutes. This operation can not cancelled. Click Next to continue. The file will be created.

8. If you did not run the wizard on the GFI MailEssentials machine itself or did not choose the option to update the GFI MailEssentials bsp file directly, you will have to copy the file to the GFI MailEssentials\data directory. Note that you will have to stop the GFI MailEssentials services and IIS admin first. After restarting the services, the new bsp file will take effect.