Abstract: Utility pattern mining is to identify the item sets with highest utilities, by considering profit, quantity, cost or other preferences. Mining high utility item set from a transaction database is finding item sets that have utility above a user-specified threshold. It extracts highly utilized items accurately, but it suffers due to the privacy breaches. Privacy preserving data mining preserves the sensitive data by modifying the structure of the data. A PPUM algorithm requires a significant amount of time to protect the privacy of data holders because they conduct database scanning operations excessively many times until all important information is hidden. The algorithm becomes inefficient in case of larger application. To overcome this drawback FPUTT(Fast Perturbation algorithm using Tree structure and Table) algorithm minimize the scanning to three times, two time for constructing tree and one time for updating On comparing FPUTT with PPUM in terms of time complexity, privacy preservation and the execution time , FPUTT is efficient than HUM-UT. The existing FPUTT algorithm uses perturbation technique for securing sensitive items which offers a low level security. The proposed work HUM-UT (High Utility itemsets Mining based on Utility Tree) algorithm handles the data streams. In order to enhance the security level, HUM-UT uses anonymization technique for protecting sensitive items The MWS (Maximal frequent pattern mining with weight conditions over data streams )Maximal frequent pattern mining rule and closed frequent pattern rule is applied, to extract maximal itemset and closed item set MMCAR(Minimal non redundant Multilevel and Cross-level Association Rules) from the given item sets. The implementation of the above techniques s increased security to item set and also extracts maximal and closed item sets from data stream. The experimental results show the proposed algorithm outperforms the existing in terms of execute time, memory of usage and number of candidates.
Keywords: Fast Perturbation, closed frequent pattern rule, High utility item set, Maximal frequent pattern mining rule.