Dissertation > Excellent graduate degree dissertation topics show

Research of Privacy Preserving Outlier Detection

Author: ZhangZuo
Tutor: XueAnRong
School: Jiangsu University
Course: Applied Computer Technology
Keywords: privacy-preserving outlier detection the semi-honest model vertical partitioned data secure multi-party computation
CLC: TP311.13
Type: Master's thesis
Year: 2010
Downloads: 95
Quote: 0
Read: Download Dissertation

Abstract


The purpose of privacy-preserving data mining is to get accurate mining results using data mining tools under the premise of protecting private data.With increasingly concerned on privacy,privacy-preserving data mining has become an important research direction.Privacy-preserving data mining is researched from secure multi-party computation and distortion.Many privacy-preserving algorithms for data mining has been proposed including association rules,classification,clustering and so on,but few algorithms are for outlier detection.Current privacy-preserving outlier detection algorithms are mainly based on distance with poor efficiency and low security because of using common secure protocol and applying central algorithms to distributed data directly. The paper combines secure multi-party computation and data distortion for outlier detection algorithms to overcome the drawbacks mentioned above.The main contributions can be summarized as follows:(1) The paper presents privacy-preserving distance based outlier detection algorithm VPPDT~k-Outlier over vertical partitioned data under semi-honest model by constructing the firstκvalue in scalar sum protocol and using data distortion technique.Theoretical analysis and experiments demonstrate that the algorithm with approximate linear communication is an efficient solution in communication and good in privacy preserving.(2) Basing on the characteristics of the density based outlier detection algorithm DBOM and data distribution,a vertical distributed density based outlier detection algorithm VDBOM is researched and proposed.Both theoretical analysis and experiment can verify that the algorithm overcome the inefficiency arising from centralized outlier detection algorithm DBOM used in distributed database.(3) Constructing and basing on the secure intersection protocol and secure comparison protocol,the paper presents privacy-preserving VDBOM algorithm under semi-honest model.Theoretical analysis demonstrated that the algorithm is effective and secure.

Related Dissertations

  1. Research on Anonymized Privacy Preserving Publishing of Data Streams,TP311.13
  2. The New Privacy Preserving Algorithms for Linear Programming and SVMS,O221.1
  3. Outlier Detection Techniques on Uncertain Moving Objects,TP311.13
  4. Research and Implementation on Lightweight Secure Data Aggregation Scheme in Wireless Sensor Networks,TP212.9
  5. Study on Secret Comparison and Its Application,TP393.08
  6. Research on Technology about Privacy-Preserving for Cross System Personalization,TP393.08
  7. Study on Privacy Preserving Classification Data Mining,TP311.13
  8. Research on Application of Secure Multi-party Computation to Key Agreement,TN918.1
  9. Research of Secure Multiparty Computation Protocals,TP393.08
  10. Research on Secure Protocols for Determing Geometric Objects Position,TP309
  11. Research on Several Secure Query Protocol of Database,TP311.13
  12. Research on Two Basic Secure Multi-party Computation Problems,TP309
  13. Research on Dynamic Adjustment Mechanism of Security and Complexity of Secure Multiparty Computation Protocol,TP309
  14. SMC for Sequencing Problem and SMC Based on Publicly Verificable Secret Sharing,TN918.1
  15. The Study on Distributed Cryptosystem and Its Application Based on Bilinear Pairing,TN918.1
  16. Secure Two-party Neural Network Computing and Learning,TP183
  17. Study on Special Secure Multi-Party Computation Protocols,TN918.1
  18. The Study on Privacy-preserving Association Rules Mining in Business Intelligence,TP311.13
  19. Research on Privacy Preserving Distributed Clustering Algorithm,TP311.13
  20. Privacy clustering - based analysis and research,TP311.13
  21. Research of Privacy-Preserving Clustering Algorithm over Distributed Data,TP311.13

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer software > Program design,software engineering > Programming > Database theory and systems
© 2012 www.DissertationTopic.Net  Mobile