Bayesian surprise is the result of mismatches between our expectations and actual results, hence the degree of surprise or anomalousness attached to a pattern will vary with respect to these differences. The implication of obtaining large surprise values identifies those patterns likely to be useful and interesting to the user.
-
Updated
Mar 5, 2024 - R