Certain Investigations on Swarm Intelligence Based Clustering Methods For Privacy Preservation Of Horizontally Partitioned Data

Abstract

In spite of the emergence of data mining as a significant technology for gaining knowledge from vast quantities of data the popularity and wide availability of data mining tools raise concerns about the privacy of the individuals Developing data mining techniques suitable for databases and maintenance of the individual privacy has become the primary aim of the research based on Privacy Preserving Data Mining PPDM Cluster analysis plays an indispensable role in accomplishing data mining applications like scientific data explorations marketing medical diagnosis and computational newlineBiology This research focuses on Privacy Preserving Clustering PPC methods for horizontally partitioned data In the previous works researchers have proposed several clustering methods such as k means clustering hierarchical k means clustering and newlinedensity based clustering for the privacy preservation of horizontally partitioned data The methods incorporate cryptographic techniques to minimize the information shared while adding little overhead to the mining task But the existing methods have some issues to be addressed including data disambiguation problem, performance of clustering only with single view point of data and applying anonymization only to Single Sensitive newlineAttributes SSA and not for Multiple Sensitive Attributes MSA Thus privacy becomes a major issue in conventional privacy preserving methods newline newline