Random Forest Algorithm Based on Linear Privacy Budget Allocation

In the era of big data with exponential growth in volume, how to reduce security issues such as leakage caused by machine learning is a hot area recent research. The existing privacy budget allocation strategies are usually only suitable for applications specific spaces and cannot meet users' personalized needs allocation. Therefore, linear strategy proposed. assigns each layer linearly increasing from root decision tree bottom adjusting coefficient or constant term. Combining this random forest algorithm, algorithm based on (DiffPRF_linear) formed. Experimental results show that proposed can realize uniform, arithmetic, geometric policy effects also achieve better classification than former, which not meets users protect private but maintains high accuracy.