The Research Of Privacy-preserving Distributed Association Rules Mining Algorithm

Data Mining is considered to extract the interested knowledge from large databases or data warehouses. With the development of database and network technology, there is growing concern for the security issues of privacy data in data mining process .Privacy-preserving data mining of association rules in distributed database is become the new direction in data mining field.The P_ODMA algorithm is proposed in this thesis that is based on the DMA algorithm and improves some aspects, such as storage space, data mining efficiency and security. Firstly, the binary coding technology is used to code all the transaction set in the database.The storage space of transaction is lowered and the speed of data mining algorithm is improved.In the aspects of data mining efficiency,the length of transaction is shortened by pruning the non-frequent itemset of transaction database .The scan time and storage space of transaction itemset are reduced by merging the iteration transaction,which reach the object of improving the data mining efficiency and reducing the complexcy of space. In the aspect of data mining privacy-preserving, the scheme of combining the RSA and HES is adopted, the security of data encryption and the efficiency of encryption algorithm are consided in this scheme, and it reaches the balance beween efficiency and security.Secondly, the filter method of association rules is improved and the new filter method based on "upgrade" is proposed. Finally, according to the property of P_ODMA algorithm, the golbal main site technolegy is adopted to reach the framework of algorithm.