K-DDD Measure and MapReduce Based Anonymity Model for Secured Privacy-Preserving Big Data Publishing

Abstract

Nowadays, big data publishing is the emerging trend since they have good potential for the decision support in the applications, such as a hospital, government, industries, etc. Existing algorithms have many problems in preserving the privacy of the data when the data is in large size. To avoid these problems, this paper introduces a novel anonymity model for the data publishing based on K-DDD measure and MapReduce. This paper presents the Duplicate-Divergence-Different properties enabled dragon Genetic (DDDG) algorithm based on the k-DDD anonymization and the dragon operator based genetic algorithm. The proposed DDDG algorithm allows the privacy preservation in the big data by modifying the MapReduce techniques with the proposed DDDG algorithm. The performance of the proposed anonymity model is analyzed with the metrics such as information loss (IL) and the classification accuracy (CA). The adult database from the UC Irvine dataset is used for the simulation. The simulation results show that the proposed DDDG algorithm achieved the lowest IL of 0.0191 and the highest CA with the value of 0.8977 than the existing algorithms for k value of 2.