A simple application of Bayes Statistics is Naive Bayes classifier. The process using document(emails) as an example is clearly documented in wiki with spam and non-spam classes.

Now the problem is reduced to figure out the likelihood of single word is spam or not and the ratio between probability of spam versus of non-spam in the whole body of emails.

Posting more concrete examples below for illustrate how this Naïve Bayes classifier works for binary data checking: