A NOVEL APPROACH FOR PRIVACY PRESERVING DATA MINING USING RANDOMIZATION

Abstract

Data mining is a knowledge extraction process from a huge amount of data. This data has many kinds such as pictorial data, analytical data, and survey data. There is information of individuals and organizations in the data to be mined. Many organizations have private and sensitive information about the individual. This sensitive information should not be publicized so that it can cause threat to the privacy of the individual. But for data mining purpose the information is needed to be publicized for various data mining tasks. Privacy preserving data mining is a recent research subject for discovering new methods so that both the purpose of data mining should be fulfilled. Privacy of the individual is protected. Scope of this project is to study the recent research work on the privacy preserving data mining and improve or propose a new naval approach for privacy preserving data mining. Among the techniques that are studied k-anonymization and randomization are used to get a new improved approach for privacy preserving data mining. In randomization method the matrix method is used to perturb the data. There are modifications that are made to the existing method. In k-anonymity generalization method is used. The combined method is proposed and analyzed so that it is efficient with respect to the concerned parameters.