An In-Depth Comparative Study of Adaptive k-Anonymity Methods for Streaming Data Privacy

The real-time data is growing extensively due to the immense use of numerous web applications, IoT devices, social media, and network-based applications. This online streaming data, characterized by its volume and velocity, is expressed as big data. While it is accessible for business analytics and research purposes, it can often sacrifice individual privacy. There are different traditional approaches, such as k-anonymity, l-diversity, and t-closeness, that exist to safeguard individual privacy by making each data record indistinguishable from at least k other records. The dynamic nature of real-time stream data makes these methods difficult to apply. However, various research shows that modifications to these methods can effectively protect individual privacy in streaming data. This paper presents a comprehensive review of k-anonymity-based techniques that adapt sliding window models, clustering approaches, and other variations to efficiently protect data privacy while maintaining k-anonymity without compromising data utility. The review discusses the challenges faced in protecting stream data privacy and concludes with research directions to enhance these methods for adaptive and scalable privacy-preserving mechanisms for streaming data.