Optimized Privacy Protection Method for Big Data Traffic

Data privacy leakages and accuracy declines are major problems in the field of big data. In this work, an adaptive space partition algorithm based on traffic history data under differential privacy is proposed to meet the privacy demands of big data traffic. This paper first uses the counting values of tree nodes in the quad-tree index structure at the first n moments by considering the spatial and temporal characteristics of traffic data and then applies the weighted moving average filter in exponential decay mode to predict the statistical values of the corresponding regions of the data to be released at the next moment. Then, an adaptive quadtree index structure based on the predicted values of each region and the heuristic judgment strategy is established by using the top-down approach and applied to the real data released at the next moment. Finally, we combine the differential privacy tree structure publishing technology and adjust the consistency constraint. Theoretical analyses and simulation experiments show that our algorithm protects data privacy and improves the accuracy of data query compared with existing algorithms.