Jan Rothkegel

Thanks for your feedback :-) Unfortunately, I don't use Outlook on my personal device anymore (because I migrated to Linux a long time ago) and therefore will most probably not implement any changes anymore.

Hi Marc, I know :-) However, I am not able to develop this any further (I don't even use Windows and Office anymore on my personal notebook). Thunderbird's spam filter does almost the same out of the box ;-) It is still a bug, but in the light of the bayes model behind it should not make such a huge difference. Best Jan

Thank you, I like to hear that my plugin helps you. The error message does not tell much. It just hides the real underlying issue (not very good practice on my side). May I ask whether you export/import your configuration from install to install?

The tool can be trained from "good" and "bad" messages. However, they must arrive in the same folder. Trying to train the model from other folders than set in your configuration will not alter the model. Please don't overestimate the impact of volume of training data. Just install and try whether it works with some (5 - 10) spam and non-spam examples.

Show statistics

Show statistics

Show statistics

Show statistics