Two-Party Privacy-Preserving Agglomerative Document Clustering

摘要

Document clustering is a powerful data mining technique to analyze the large amount of documents and structure large sets of text or hypertext documents. Many organizations or companies want to share their documents in a similar theme to get the joint benefits. However, it also brings the problem of sensitive information leakage without consideration of privacy. In this paper, we propose a cryptography-based framework to do the privacy-preserving document clustering among the users under the distributed environment: two parties, each having his private documents, want to collaboratively execute agglomerative document clustering without disclosing their private contents.