Hybrid privacy preservation model for big data publishing on cloud

Abstract

Cloud computing has gained popularity since the user can share a large amount of data through the cloud servers. Cloud servers provide services to a large number of users simultaneously. The third party users access the information provided by the various users of the cloud network. While sharing the information of the licensed user to the third party, user privacy and the utility of the information need to be maintained. Various techniques have been introduced for increasing the privacy of the information. This work introduces an anonymisation based technique for the privacy preservation of the data before publishing it in the cloud environment. This paper uses the k-anonymisation criteria for duplicating k records in the information database. The proposed preservation model uses the dragonfly (DF) algorithm for achieving the k-anonymisation database. The proposed work defines a fitness function for achieving maximum privacy and utility during the process. The performance of the proposed work is analysed with the metrics such as classification accuracy and information loss. Various comparative methods are used for analysing the performance of the proposed work. From the simulation results, the proposed privacy preservation model with the k-anonymisation and the DF has the better values for the classification accuracy and the information loss.