STAMP: On Discovery of Statistically Important Pattern Repeats in Long Sequential Data