Privacy Preservation using Utility Itemset Mining via Transaction Splitting

Nowdays, businesses are evolving. For having business people needs to deal with much amount of data and this data needs to be delicate and confidential. So, to secure and preserve our data there are plenty of technologies used one of them is Data Mining. Data Mining is the technique in which it tries to find out interesting patterns or knowledge from database such as association or correlation etc. Frequent Itemset Mining is the vital problem in data mining. The frequent can contains valuable and research purpose. Frequent itemsets are items or patterns like itemset, substructures or subsequences that occurs frequently in transaction. There are many Frequent Itemset Mining Algorithms to find out frequent itemset used such as Apriori, FP-growth, Elcat. But this all only consider frequent itemset don’t consider useful or high utility itemset from large transaction. For, Privacy and security purpose here use differentially Utility Itemset Mining using transaction splitting and Utility Pattern Growth algorithm. It consists of two phase that are preprocessing phase and mining phase. In preprocessing phase, to boost the value of utility and privacy advance smart splitting method is used to transform database. Preprocessing phase is performed only once in given database, Mining phase, execute runtime estimation and dynamic reduction. To sustain information loss by smart splitting, it used runtime estimation which estimates actual support of itemsets from original database. For privacy preservation, here added noise in the database to reduce the noise it apply forward dynamic method. For mining high utility itemset and privacy preservation proposed new algorithm called Utility Pattern Growth.