Privacy in Data Mining

1
Privacy in Data Mining

• Presented by
• Kalyan K Beemanapalli Vamshi
  Kodithava
• Graduate Student
  Graduate Student
• Department of CS
  Department of CS

2
Outline

• Definitions, Introduction, Importance and
  Relevance to the Class
• Key Issues and Key Results
• Techniques Developed
• Open Issues/ Research Directions
• References
• Conclusion

3
Definitions..

• Data Mining Data mining - or knowledge discovery
  in data bases (KDD) is the nontrivial process
  of identifying valid, novel, potentially useful,
  and ultimately understandable patterns in data
• Privacy Individuals desire and ability to keep
  certain information about themselves hidden from
  others. What is Corporate Privacy?
• Relationship The primary goal of data mining is
  to extract hidden relationships and patterns
  among data items. Hence the privacy concerns of
  data mining

4
Importance.

• Example from Health Informatics of privacy
  breaches due to data mining
• The Database Inference Problem
• Data mining makes the Inference problem quite
  dangerous
• Hence, data mining algorithms are being revisited
  from the angle of privacy, security and Civil
  Liberties
• Relevance to Class Before applying Data Mining
  Techniques it is important to know its positive
  and negative implications

5
Outline

• Definitions, Introduction, Importance and
  Relevance to the Class
• Key Issues and Key Results
• Techniques Developed
• Open Issues/ Research Directions
• References
• Conclusion

6
Privacy Preserving Data Mining

• Approaches adopted for preserving privacy
• Data Distribution
• Data Modification
• Data Mining Algorithm
• Data or rule hiding
• Privacy Preservation
• Concerns Performance, Scalability, Data Utility,
  Level of Uncertainty, Resistance
• Very thin line between utilizing the power of
  data mining techniques and preserving privacy.

7
Heuristic-Based Techniques

• Association Rule Confusion
• Classification Rule Confusion
• Privacy Preserving Clustering
• Cryptography-Based Techniques
• Reconstruction-Based Techniques
• These techniques are combined with various data
  modification techniques and different data
  distribution techniques and numerous algorithms
  have been developed

8
Reconstruction-Based Techniques

• Outcome of the work done by Rakesh Agrawal and
  Ramakrishnan Srikant at IBM Research, Almaden
• Step 1Creating randomized data sample by data
  perturbation of individual data records
• Step 2 Reconstruct distributions, not values in
  individual records.
• Step 3 By using the reconstructed distributions,
  build the decision tree classifier
• For Reconstruction they used the Bayesian
  approach and proposed three algorithms for
  building decision trees that rely on
  reconstructed distributions

9
Accuracy Analysis
10
Outline

• Definitions, Introduction, Importance and
  Relevance to the Class
• Key Issues and Key Results
• Techniques Developed
• Open Issues/ Research Directions
• References
• Conclusion

11
Open Issues/ Research Directions

• Data Warehousing and the Inference problem
• Preparing Perturbed Databases for a combination
  of Algorithms
• Social Effects. Work with Social Scientists to
  preserve privacy over cultures
• Formulating legal rules and developing data
  mining algorithms accordingly
• Privacy Inference Controller
• Quantifying Privacy?
• How about formulating something similar to ACID
  properties for Perturbed Databases?

12
References

• Security and Privacy Implications of Data Mining,
  1996, Chris Clifton and Don Marks
• Defining privacy for Data Mining , Chris Clifton,
  Murat Kantarcioglu and Jaideep Vaidya, Purdue
  University
• Data Mining, National Security, Privacy and Civil
  Liberties, Bhavani Thuraisingham, The National
  Science Foundation
• A Framework for Privacy Preserving Classification
  in Data Mining, 2004 Md.Zahidul Islam and
  Ljiljana Brankovic
• Privacy Preserving Mining of Association Rules,
  2002, Alexandre Evfimlevski, Ramakrishnan
  Srikant, Rakesh Agrawal and Johannes Gehrke, IBM
  Almaden Research Center

13
References Continued

• Privacy Preserving Data Mining, 2000, Rakesh
  Agrawal and Ramakrishnan Srikanth, IBM Research,
  Almaden
• Privacy Preserving Data Mining, Advances in
  Crptology, 2000, Y.Lindell and Benny Pinkas.
• Detecting Privacy and Ethical Sensitivity in Data
  Mining Results, 2004, Peter Fule and John Roddick
• Limiting Privacy Breaches in Privacy Preserving
  Data Mining, 2003, Alexandre Evvfimievski,
  Johannes Gehrke and Ramakrishnan Srikant
• State-of-the-art in Privacy Preserving Data
  Mining, Vassilios S.Verykios, Elisa Bertino, Igor
  Nai Fovino, Provenza, Yucel Saygin and Yannis
  Theodoridis

14
Conclusion

• Presented an overview and brief insight into the
  new research in the area of data mining
• Statistical Databases were the first to think
  about privacy issues when data is being analyzed
• Techniques and approaches used by researchers for
  preserving privacy while mining the data
• Various open issues which have to be addressed
  and the potential research directions
• Privacy, Secrecy, Ethical Sensitivity , National
  Security Civil Liberties.

15
Conclusion.Continued

• There is so much to talk about this issue of
  privacy in general and specifically related to
  data mining.
• Data Mining can be used to its fullest ability
  only if researchers address the problem of
  privacy and develop techniques in this direction

16
Queries
Thank You.