An Efficient Privacy Preserving Protocol for Dynamic Continuous Data Collection

Abstract Past and ongoing decades have witnessed significant uplift in data generation due to ever growing sources of data. Collection and aggradation of such huge data have triggered serious concerns on privacy of data-owners’ sensitive information. Catering this, several existing anonymization models proffer privacy-preserving data collection. However, the models put-forth either strict or unrealistic assumptions regarding leaders’ selection (the concept of first and last leaders in data collection process). In this paper, we have identified and formally defined a privacy attack, Leader Collusion Attack (LCA) ; where first and second leaders may collude to breech individuals’ privacy during data collection process. In this regard, we have proposed a novel k -anonymity based dynamic data collection protocol (presented single leader election) to mitigate LCA . Moreover, we have formally modelled and analysed the proposed protocol through HLPNs and demonstrated the mitigation of LCA . Experimentations on real-world datasets advocate the outperformance of our protocol over existing model in terms of better utility and privacy levels.