Wireless Big Data for Differential Seclusion Preserving
With the popularity of smart devices and the widespread use of machine learning methods, smart edges have become the mainstream of dealing with wireless big data. When smart edges use machine learning models some models may unintentionally store a small portion of the training data with sensitive records. To solve this privacy issue, in this paper, we proposed and implemented a machine learning strategy for smart edges using differential privacy. Our attention has been focused on privacy protection in training datasets in wireless big data scenario. Privacy protection adds Laplace mechanisms, and designed three different algorithms which satisfied differential privacy. This project introduces a privacy preserving approach that can be applied to decision tree learning, without connected loss of accuracy. Meanwhile, an accurate analysis can be built directly from those unreal data sets. This can be applied directly to the data storage as soon as the first sample is collected. The Relevant Columns Values Swapping approach is compatible with other privacy preserving approaches, such as without cryptography, for extra protection.