Please use this identifier to cite or link to this item:
|Title:||Secure data mining of outsourced data||Authors:||Liu, Fang||Keywords:||DRNTU::Engineering::Computer science and engineering||Issue Date:||2016||Source:||Liu, F. (2016). Secure data mining of outsourced data. Doctoral thesis, Nanyang Technological University, Singapore.||Abstract:||Organizations and individuals nowadays are more and more willing to outsource their data to save storage and management costs, especially with the push for cloud computing which is service-oriented and offers both storage and computation scalability. However, the data, once being released to a server, is no longer under its owner’s control, and its privacy and security herein become a primary concern. To this end, users usually encrypt the private data before outsourcing it, which however makes conventional data retrieve, sharing, and analysis services be very thorny and challenging as data is both big and encrypted. Under such new circumstance, diverse secure building blocks and some more complex secure data mining techniques should be considered for secure analytical computations and knowledge discovery on outsourced databases. In this thesis, we aim at investigating various secure data mining algorithms for the cloud platform where data is centralized and encrypted. To enhance the security, we select suitable cryptographic techniques to protect user’s privacy and to allow a cloud server to manipulate encrypted data. According to our objectives, we first discuss and analyze several secure issues caused by outsourcing data to the cloud, such as query executing techniques, multiple user key management, correctness and integrity verifying, privacy-preserving data mining algorithms, and so on. Second, we design some basic secure building blocks for the cloud platform, including secure set intersection and secure scalar product. Third, based on such secure building blocks, we formally develop three secure data mining protocols to perform following data mining algorithms: association rule mining, gradient descent algorithm, and SVM classification. Finally, the thesis makes the conclusion and the prospect of further research directions.||URI:||http://hdl.handle.net/10356/67021||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||SCSE Theses|
Page view(s) 20146
checked on Sep 28, 2020
checked on Sep 28, 2020
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.