An Adaptive Hidden Markov Model with CCA for Privacy Preserving of Correlated Big Data

Abstract

Due to technological advancements and increase in the use of smart devices, huge amount of data is generated and has open access to various social media servers all around the world. Most of the social media providers seldom care on security and or preservation of private data. One of the greatest challenges that prevail due to the existence of correlated information is privacy preserving data mining. Many research methodologies have been proposed yet maintaining the privacy in social network is a complex process. In this proposed method, a novel methodology for preserving the privacy of Correlated big data using various techniques. The proposed method consists of three main processes they are, correlated big data identification, correlated big data analysis and correlated iteration mechanism. In first process an adaptive Hidden Markov Model (AHMM)used for identifying the Correlated big data present in the datasets. Then in the second process, using the canonical correlation analysis (CCA) the correlation matrix is find out for sensitivity measure. In last process, to answer the large group of queries designed a correlated iteration mechanism. Thus implemented the proposed system and the implementation results are compared with the conventional techniques. Ultimately the proposed method suggests that the performance is better for the privacy preserving of correlated dataset.