Towards a Framework for Privacy Preserving Medical Data Mining Based on Standard Medical Classifications
Privacy-preserving data mining often focuses on data alteration but may bias data patterns interpretation and does not offer different levels of access to patterns according to their use. This paper addresses data mining as a prediction tool and proposes to offer several levels of access to data patterns according to users’ trustworthiness. The grounding intuition is that patterns’ predictive value depends on their precision that should thus vary according to their use. The following problem is considered: a medical data holder wants to disclose data or data patterns and still control the meaning of the disclosed patterns or of the patterns that may be mined out of the released dataset. To tackle this issue, we propose a framework compliant with existing data mining techniques by modeling trust in terms of data precision and generalising data according to standard medical classifications.
KeywordsPrivacy Data Pattern Hidding Data Generalisation
Unable to display preview. Download preview PDF.
- 1.Agrawal, R., Srikant, R.: Fast algorithms for mining association rules in large databases. In: Proceedings of the 20th International Conference on Very Large Data Bases, pp. 487–499. Morgan Kaufmann, Santiago de Chile (1994)Google Scholar
- 5.Freund, J., Comaniciu, D., Ioannis, Y., Liu, P., McClatchey, R., Morley-Fletcher, E., Pennec, X., Pongiglione, G., Zhou, X.: Health-e-child: An integrated biomedical platform for grid-based paediatric applications. CoRR abs/cs/0603036 (2006) Google Scholar
- 10.Sweeney, L., Sweeney, L.: Achieving k-anonymity privacy protection using generalization and suppression. International Journal on Uncertainty, Fuzziness and Knowledge-Based Systems 10 (2002) Google Scholar
- 12.Wang, K.: Bottom-up generalization: a data mining solution to privacy protection. In. In: ICDM, pp. 249–256 (2004)Google Scholar