Frequent Itemset Mining using PFP-Growth via Smart Splitting

Frequent itemset mining has been growing interest in designing differentially private data mining algorithms. Mining frequent itemset is one of the important issues in the domain of data mining. In proposed system, FIM i.e. frequent itemset mining algorithm is proposed which not only mines high data utility, degree of privacy but it also offers high time efficiency. To achieve privacy, private FIM algorithm is proposed. It is based on FP-growth algorithm hence it is known as, PFP-growth algorithm. There are two phases involved in PFP-growth algorithm such as, preprocessing phase and mining phase. Smart splitting method is used in preprocessing phase for transformation algorithm which enhances the utility and trade off. In mining phase, runtime estimation method is used for estimation of actual support of itemset in the original database. To reduce the noise added during mining process for guarantee of privacy is reduced using dynamic reduction method with map-reduced framework. With experimental results proposed techniques outputs better efficiency in terms of time and memory.