プライバシ保護したK-meansクラスタリング対話型プロトコル

概要

論文の詳細を見る
K-means clustering is a powerful and frequently used technique of Data mining, the traditional techniques which require access to precise information in individual records will cause privacy problem. For this concern, privacy-preserving k-means technique is needed. In this paper, we consider a scenario in which two parties owning confidential databases wish to run a k-means clustering algorithm to do data mining task on the union of their databases, without revealing any unnecessary information under the semi-honest model. We focus on the problem of k-means clustering algorithm and give out a proposal which can deal with both numeric data and non-numeric data to execute the k-means clustering algorithm while preserving the privacy. The former research results often employ either Trust Third Party or cryptographic methods with too much complexity and cannot process the non-numeric attributes. Our solution is using the secure multi-party computation techniques to construct privacy-preserving k-means clustering protocols with less communication and computational complexity, and we show our solution can process the non-numeric attributes over two parties' databases.
社団法人電子情報通信学会の論文
2005-09-09