How SpamAssassin defines BAYES_99 ? How to avoid it?

SpamAssassin marks my email with reason BAYES_99 , which means the email is 99% ~ 100% spam email. What is the mechanism of SpamAssassin to det开发者_JAVA技巧ect this rule (by keyword?), and how can I avoid hitting this rule?

I found this because I was using SendGrid for transactional mail, and sendgrid was sending MY mail from a server that had been blackholed for sending SOMEONE ELSE'S spam. Because the IP was blackholed, my emails were marked spam. To diagnose:

You can see if your domain or IP has been blacklisted here: https://mxtoolbox.com/blacklists.aspx

The fix was to switch to a different email transport provider.

BAYES_99 is a statistical component of SpamAssasin. It is using some sort/variation of Naive Bayes which has (to keep it simple) some sort of probability saying that some word/token is likely to be spam or not to be spam.

In "math" these could be expressed like this:

P(Class_Spam | Email)
P(NotClass_Spam | Email)

Transformations lead to e.g. this equation:

However there a lots of different representations of bayes classifiers. With different optimizations and statistical models.

You may disable it via spamassasin configuration. This question is not well placed on stackoverflow. Superuser would be the better fit though.