Thanks for this - I have implemented this in my project. But it seems that there are some false-positives, so I decided not to remove the Spam - Messages but to count them for each author and calculate a 'spam-score' based on Spam and no spam messages. However - how could I add a function to add spam messages to the AI-Base? I mean, to enhance the basis of HAM / SPAM over time by classifying messages?
But it seems that there are some false-positives, so I decided not to remove the Spam - Messages but to count them for each author and calculate a 'spam-score' based on Spam and no spam messages.
However - how could I add a function to add spam messages to the AI-Base? I mean, to enhance the basis of HAM / SPAM over time by classifying messages?
Regards,
Thorsten