An algorithm to discover the approximate probabilistic frequent itemsets with sampling method

摘要

Probabilistic frequent itemset mining is a challenging problem when mining uncertain databases. We address this problem and present an algorithm named APFIMSample. It presents a probabilistic support estimating method to reduce the computing cost, which together with a sampling technique, can significantly improve the performance. We evaluate our algorithm over two datasets, and the experimental studies show that the APFIMSample algorithm is effective.