Privacy Preserving Parallel Clustering Based Anonymization for Big Data Using MapReduce Framework

Big data refers to a massive volume of data collected from heterogeneous data sources including data collected from Internet of Things (IoT) devices. Big data analytics is playing a crucial role in extracting patterns that would benefit efficient and effective decision making. Processing this massive volume of data poses several critical issues such as scalability, security and privacy. To preserve data privacy, numerous privacy-preserving data mining and publishing techniques exist. Data anonymization utilizing data mining techniques for preserving an individual’s privacy is a promising approach to prevent the data against identity disclosure. In this paper, a Parallel Clustering based Anonymization Algorithm (PCAA) is proposed, and the results prove that the algorithm is scalable and also achieves a better tradeoff between privacy and utility. The MapReduce framework is used to parallelize the anonymization process for handling a huge volume of data. The algorithm performs well in terms of classification accuracy, F-measure, and Kullback–Leibler divergence metrics. Moreover, the big data generated from heterogeneous data sources are efficiently protected to meet the ever-growing requirements of the application.