Algorithms for Frequent Itemset Mining and Database Sanitization

Data mining techniques have been widely applied in numerous areas and represent an important field of research. In Chapter 1, the research motivation, objectives and contributions are introduced. Chapter 2 introduces background work on data mining, share mining, utility mining, and privacy-preserving data mining. Chapter 3 describes the proposed NFP-growth method for discovering frequent itemsets. Chapters 4 through 6 explain several novel fast algorithms for share mining --- including FSM, EFSM, SuFSM, ShFSM, and DCG --- to efficiently generate all share- frequent itemsets. Furthermore, Chapter 7 presents the Isolated Items Discarding Strategy (IIDS), which can be applied to any existing level-wise share mining or utility mining method to reduce candidates and to improve its performance. Next, Chapter 8 introduces the proposed Maximum Item Conflict First (MICF) algorithm, which has a low sanitization rate and achieves a low misses cost, for hiding all restrictive itemsets. At the end of Chapters 3 through 8, the experimental results and evaluates the performance of the proposed algorithms are provided. Finally, Chapter 9 draws a summary of the dissertation.