A Privacy Enhanced Data Aggregation Model
Shu Qin Ren, Khin Mi Mi Aung, Jong Sou Park · 2010
This paper proposes an enhanced privacy-preserving data aggregation scheme, which balances the onerous task of extracting reasonable data value and preserving data privacy even with incomplete or malicious data presentence. We propose an innovative encryption algorithm to preserve data privacy while it can provide secure data comparison between the encrypted data. Furthermore, we define a robust and efficient aggregation operator to fuse the encrypted data without decryption by secure data comparison and density based data mining. The proposed aggregation scheme can remove both potentially malicious and redundant data before aggregation so that it can provide a robust aggregation result without scarifying data privacy. We also discuss the scheme performance in terms of aggregation accuracy, distribution recovery ability and aggregation efficiency. The experiment results show that this scheme can give reasonable aggregation values, recover the data distribution well even under 50% malicious readings, much more robust than the commonly used aggregation while it has good aggregation efficiency with above 80% redundant data removal.