Privacy-Preserving Classification for Big-Data Driven Network Traffic. Protecting sensitive information in large network traffic flows while ensuring data usability for classification emerges as a critical problem of increasing significance. Existing techniques do not work on highly heterogeneous traffic from big-data applications for both privacy protection and classification (such as port-based and load- based methods). This project investigates new theories, methods and techniques for solving ....Privacy-Preserving Classification for Big-Data Driven Network Traffic. Protecting sensitive information in large network traffic flows while ensuring data usability for classification emerges as a critical problem of increasing significance. Existing techniques do not work on highly heterogeneous traffic from big-data applications for both privacy protection and classification (such as port-based and load- based methods). This project investigates new theories, methods and techniques for solving this problem. It proposes to develop a set of effective methods for privacy-preserving data publication through combining randomisation with anonymisation, and for classifying the published data through uncertainty leveraging by probabilistic reasoning and accuracy lifting by inter-flow correlation analysis and active learning.Read moreRead less