A sampling-based method for mining frequent patterns from databases

Mining frequent item sets (frequent patterns) in transaction databases is a well known problem in data mining research. This work proposes a sampling-based method to find frequent patterns. The proposed method contains three phases. In the first phase, we draw a small sample of data to estimate the set of frequent patterns, denoted as F^S. The second phase computes the actual supports of the patterns in F^S as well as identifies a subset of patterns in F^S that need to be further examined in the next phase. Finally, the third phase explores this set and finds all missing frequent patterns. The empirical results show that our algorithm is efficient, about two or three times faster than the well-known FP-growth algorithm.