Genetic Algorithm Based Confidential Data Protection in Privacy Preserving Data Mining

Abstract: Nowadays data sharing becomes obvious for successful completion of several tasks. In many places, data sharing has created lot of problems i.e. data modification, data mishandling and data misinterpretation. If we wish to share the datasets to others, first we have to verify whether the dataset has any sensitive or confidential data, if yes, then it is necessary and essential to protect that confidential data and then it can be shared to others. Privacy preserving data mining helps to perform the data mining tasks in a secured manner. Many privacy techniques and algorithms have been proposed by many researchers. This research work Genetic Algorithm based Masking Technique (GAMT) mainly concentrated on protecting confidential numerical attributes in a dataset using the genetic algorithm. Genetic algorithm concept is used for modifying the data items of the confidential attribute and to generate new data values for those attributes. The performance of Genetic Algorithm based Masking Technique is compared with the existing technique, named as additive noise. Results showed that GAMT has produced better results.

Keywords: Data mining, Privacy, Genetic algorithms, Additive noise, X-means, Filtered and Simple EM Clustering.