PRIVACY PRESERVING USING ENSEMBLE CLASSIFICATION FOR HEART DISEASE DATA SETS

Abstract

In this modern era of thriving technology, the data being gathered through way of private in addition to public businesses is increasing each day. But in recent times people are more worried about their data and privateness being preserved at the same time as use of their data in other analysis purpose. Thus Privacy-Preserving Data Mining (PPDM) method has been proposed to permit the extraction of understanding from data at the same time as keeping the privateness of people. The primary purpose of our project is on preserving privacy for healthcare records as privateness lacks in Medical data. Privacy-Preserving Data Mining (PPDM) offers with shielding the privacy of individual’s data or sensitive data without the utility of data. Therefore the strategies like anonymization, randomization are used to attain the intention. However, unfortunately anonymization results in certain level of information loss while preserving privacy. In order to overcome this problem, perturbation technique is carried out. Our challenge initiates with cleaning and preprocessing followed by ensemble classification and proceeded with perturbation to attain the goal. This method focuses on preserving privacy by perturbing the sensitive attributes in the Medical data without causing loss to the information in the process