Privacy-Preserving Mechanism for Data Analytics

This paper proposed a mechanism to maintain the data subject's privacy while performing analytics on electricity billing data. First, this paper implemented privacy-preserving mechanisms such as generalisation, group shuffling, suppression and full masking in a mocked electricity billing dataset. This paper then calculates the data utility metric to prove that the data is adequately preserved. Finally, the data utility of the preserved data is evaluated to ensure the preserved data is still usable to perform analytics tasks. Among the three mechanisms examined in this article, the group shuffling mechanism achieved the most outstanding visibility; hence, it is the most suitable mechanism to be used in data analytics. Apart from that, group shuffling generates a very little loss of information.