Research On Maximum Frequent Itemsets Based On Improved FP-tree

Data Mining is one of the most active research fields, especially in the fields of artificial intelligence and database. Data Mining is a kind of process which can discover potential useful knowledge from massive data. The association rule mining is a main research aspect of data mining, and the discovery of the frequent itemsets is a key problem of the association rule mining.FP-growth algorithm is one of the most efficient frequent pattern mining methods. However, FP-growth algorithm must generate a huge number of conditional FP-trees recursively in processes of mining maximum frequent, so the efficiency of it is unsatisfactory. In this paper, an efficient mining maximum frequent algorithm is proposed, it unifies the improvement FP-tree, the FP-tree is a one-way tree and there is no pointers to point its children in each node, so it can save the massive memories space, by introducing set of item sequences and its operators the algorithm doesn't generate conditional FP-tree or a large number of number of candidate sets in mining process, which can conveniently get all maximum frequent itemsets. The example analysis is feasibility and effectiveness of the algorithm.The sensitive association rule hiding is a very important issue in the data mining domain, the goal of which is to maintain other characteristics of the primitive data set under the condition that the sensitive rule should not be discovered. The original method based on the changes of original transaction data will generate massive I/O operations. In order to enhance the protect extensions of sensitive data and to ensure the accuracy of the mining results, uses the FP-tree to store all of the information related to transaction database and proposes an effective method of hiding the sensitivity association rule: first of all, to mine rapidly the maximum of frequent item sets; second of all, to set down the sensitive association rules; then, to delete the frequent item sets which support the sensitivity of association rules, and to update FP-tree accordingly; thus, by reversely mining the updated FP-tree, a new transaction database is generated, in which sensitive association rules are not contained. The examples and the theoretical analysis show that this method is right and efficient.