Bayesian statistics

Experimental statistics in which the assumptions about parameters are continually revised in light of new data by using a weighted average of the previous assumption (called a prior). The relative weight assigned to a prior depends on its variance over time. See also Bayes' theorem.