Input Privacy Research Articles

Sensitive Items in Privacy Preserving — Association Rule Mining

Privacy-preserving has recently been proposed in response to the concerns of preserving personal or sensible information derived from data-mining algorithms. For example, through data-mining, sensible information such as private information or patterns may be inferred from non-sensible information or unclassified data. As large repositories of data contain confidential rules that must be protected before published, association rule hiding becomes one of important privacy preserving data-mining problems. There have been two types of privacy concerning data-mining. Output privacy tries to hide the mining results by minimally altering the data. Input privacy tries to manipulate the data so that the mining result is not affected or minimally affected. For some applications certain sensitive predictive rules are hidden that contain given sensitive items. To identify the sensitive items an algorithm SENSITEM is proposed. The results of the work have been given.

Hiding collaborative recommendation association rules

The concept of Privacy-Preserving has recently been proposed in response to the concerns of preserving personal or sensible information derived from data mining algorithms. For example, through data mining, sensible information such as private information or patterns may be inferred from non-sensible information or unclassified data. There have been two types of privacy concerning data mining. Output privacy tries to hide the mining results by minimally altering the data. Input privacy tries to manipulate the data so that the mining result is not affected or minimally affected. For output privacy in hiding association rules, current approaches require hidden rules or patterns to be given in advance [10, 18---21, 24, 27]. This selection of rules would require data mining process to be executed first. Based on the discovered rules and privacy requirements, hidden rules or patterns are then selected manually. However, for some applications, we are interested in hiding certain constrained classes of association rules such as collaborative recommendation association rules [15, 22]. To hide such rules, the pre-process of finding these hidden rules can be integrated into the hiding process as long as the recommended items are given. In this work, we propose two algorithms, DCIS (Decrease Confidence by Increase Support) and DCDS (Decrease Confidence by Decrease Support), to automatically hiding collaborative recommendation association rules without pre-mining and selection of hidden rules. Examples illustrating the proposed algorithms are given. Numerical simulations are performed to show the various effects of the algorithms. Recommendations of appropriate usage of the proposed algorithms based on the characteristics of databases are reported.

Hiding informative association rule sets

Privacy-preserving data mining, is a novel research direction in data mining and statistical databases, where data mining algorithms are analyzed for the side effects they incur in data privacy [Verykios, V., Bertino, E., Fovino, I. G., Provenza, L. P., Saygin, Y., & Theodoridis, Y. (2004). State-of-the-art in privacy preserving data mining. SIGMOD Record 33 (1), 50–57, March 2004]. For example, through data mining, one is able to infer sensitive information, including personal information or even patterns, from non-sensitive information or unclassified data. There have been two types of privacy concerning data mining. The first type of privacy, called output privacy, is that the data is minimally altered so that the mining result will not disclose certain privacy. The second type of privacy, called input privacy, is that the data is manipulated so that the mining result is not affected or minimally affected. In output privacy of hiding association rules, current approaches require hidden rules or patterns been given in advance [Dasseni, E., Verykios, V., Elmagarmid, A., & Bertino, E. (2001). Hiding association rules by using confidence and support. In Proceedings of 4th information hiding workshop, Pittsburgh, PA (pp. 369–383); Oliveira, S., & Zaiane, O. (2002). Privacy preserving frequent itemset mining. In Proceedings of IEEE international conference on data mining, November (pp. 43–54); Oliveira, S., & Zaiane, O. (2003). Algorithms for balancing privacy and knowledge discovery in association rule mining. In Proceedings of 7th international database engineering and applications symposium (IDEAS03), Hong Kong, July ; Oliveira, S., & Zaiane, O. (2003). Protecting sensitive knowledge by data sanitization. In Proceedings of IEEE international conference on data mining, November 2003 ; Saygin, Y., Verykios, V., & Clifton, C. (2001). Using unknowns to prevent discovery of association rules. SIGMOND Record 30 (4), 45–54; Verykios, V., Elmagarmid, A., Bertino, E., Saygin, Y., & Dasseni, E. (2004). Association rules hiding. IEEE Transactions on Knowledge and Data Engineering 16 (4), 434–447]. This selection of rules would require data mining process to be executed first. Based on the discovered rules and privacy requirements, hidden rules or patterns are then selected manually. However, for some applications, we are interested in hiding certain constrained classes of association rules such as informative association rule sets [Li J., Shen H., & Topor R. (2001). Mining the smallest association rule set for predictions. In Proceedings of the 2001 IEEE international conference on data mining (pp. 361–368)]. To hide such rule sets, the pre-process of finding these hidden rules can be integrated into the hiding process as long as the predicting items are given. In this work, we propose two algorithms, ISL (Increase Support of LHS) and DSR (Decrease Support of RHS), to automatically hiding informative association rule sets without pre-mining and selection of hidden rules. Examples illustrating the proposed algorithms are given. Numerical experiments are performed to show the various effects of the algorithms. Recommendations of appropriate usage of the proposed algorithms based on the characteristics of databases are reported.

Secure Multi-Party Computation without Agreement

It has recently been shown that authenticated Byzantine agreement, in which more than a third of the parties are corrupted, cannot be securely realized under concurrent or parallel (stateless) composition. This result puts into question any usage of authenticated Byzantine agreement in a setting where many executions take place. In particular, this is true for the whole body of work of secure multi-party protocols in the case that a third or more of the parties are corrupted. This is because these protocols strongly rely on the extensive use of a broadcast channel, which is in turn realized using authenticated Byzantine agreement. We remark that it was accepted folklore that the use of a broadcast channel (or authenticated Byzantine agreement) is actually essential for achieving meaningful secure multi-party computation whenever a third or more of the parties are corrupted. In this paper we show that this folklore is false. We present a mild relaxation of the definition of secure computation allowing abort. Our new definition captures all the central security issues of secure computation, including privacy, correctness and independence of inputs. However, the novelty of the definition is in decoupling the issue of agreement from these issues. We then show that this relaxation suffices for achieving secure computation in a point-to-point network. That is, we show that secure multi-party computation for this definition can be achieved for any number of corrupted parties and without a broadcast channel (or trusted pre-processing phase as required for running authenticated Byzantine agreement). Furthermore, this is achieved by just replacing the broadcast channel in known protocols with a very simple and efficient echo-broadcast protocol. An important corollary of our result is the ability to obtain multi-party protocols that remain secure under composition, without assuming a broadcast channel.