ਅਕਤੂ 16, 2025 ਤੇ Sebastian (Developer) ਵੱਲੋਂ · ਪਰਮਾਲਿੰਕ

Thank you for your feedback, and I'm sorry you're running into this issue. This is a classic behavior for a new Bayesian filter. It might seem like it's tagging emails randomly, but it's actually making decisions based on the few words it has learned so far. Once it has more training data, its accuracy improves dramatically.

Here is a quick 3-step guide to fix it:

Reset the faulty tag: In the addon's settings, find the tag that is being misapplied and click "Delete Training Data" to start fresh.

Re-train with good & bad examples: Use "Learn as..." for about 10-15 correct emails. Most importantly, use "Unlearn as..." for another 10-15 incorrect emails to teach the filter the difference.

Check the threshold: In the settings, ensure the threshold for the tag is set to a high value, like 95% or more.

This balanced training is key to accuracy and should solve the problem completely. Thanks for giving PrioMailbox a try!