Anonymization based privacy preservation over cloud data using incremental clustering and map reduce approach on big data

Abstract

newline quotThe major challenge of privacy preservation over the cloud is handling the incremental data because the cloud data may be updated continuously. In this thesis, an incremental clustering technique called K-Means incremental clustering is used over incremental cloud data to efficiently cluster and update huge-volume of incremental data sets. Given a set of records as input choose some privacy sensitive attributes as quasi-identifiers for anonymization. After anonymization, the records are clustered based on K-Means incremental clustering technique, check the k-anonymity constraint, information loss for every newlinecluster and modify the cluster based on the K-Anonymity constraint. The proposed technique is compared with the existing Xuyun Zhang et al. s technique based on the updating time by adding different number of records. The time taken to update is compared for different k values. Proposed system performs well when numbers of records are large. Evaluation results have demonstrated that the efficiency of privacy preservation on large volume of incremental datasets can be improved significantly over existing approaches. newlinequot newline