Description

Privacy preservation is important for machine learning and data mining. The aim of our project is to be preserve privacy before the data are published. But the measures designed to protect private information often result in a trade-off: reduced utility of the training samples. Our paper introduces a privacy preserving approach that can be applied to decision tree learning, without loss of accuracy. It describes an approach to the preservation of the privacy of collected data samples in cases where information from the sample database has been partially lost. This novel approach can be applied directly to the data storage as soon as the first sample is collected & converts the original sample data sets into a group of unreal data sets. After the entire group of unrealized format is obtained, we can reconstruct the data, & get the original data samples. At the same time, using c4.5 algorithm an effective decision tree can be built directly from those unreal data sets. Using the approach the privacy is achieved for each & every data samples. This method provides high security compare to previous approaches.