Recently, large amount of data is collected by various organizations. Generally, data consists of various attributes such as name, address, medical term, etc. Related to the same person, different organizations often possess data with different attributes. If we can integrate data kept in different organization related to the same person without violating privacy, detailed analyzes such as cause investigation or relations among attributes could be realized. In such a scenario, we do not need personal information while it should be protected securely. Importantly, the data exactly integrates data associated with the same person. In this paper, we classify attributes in data into three of matching attributes, analyzing attributes, and others.Then, we propose a privacy preserving data integration protocol while handling data privacy appropriately according to classification of matching, analyzing attributes, and others. In addition, we propose a protocol using T-MPSI, of which integration condition is more practical, and evaluate the practicality of our protocol in terms of the amount of computation and communication for each entity.

Top