Abstrakt

Improving Privacy And Data Utility For High- Dimensional Data By Using Anonymization Technique

P.Nithya, V.Karpagam

Privacy Preserving is one of the significant methods in data mining to hide the sensitive information. Anonymization techniques like generalization and bucketization have been used for privacy preserving. The main problem with generalization is it is not applicable for high-dimensional data and bucketization technique does not avoid membership disclosure. Slicing is one of the novel techniques in which the data is partitioned horizontally and vertically. This reduces the dimensionality of the data and it is able to handle high dimensional data better when compared to generalization and bucketization. In slicing, every attribute is in exactly one column. It provides better privacy but there is loss of data utility. Overlapping slicing is a novel technique that allows duplicating an attribute in more than one column so that more attribute correlations is achieved for better data utility. For protecting membership information, a more effectual tuple grouping algorithm is proposed and continuous attributes are handled. Further to improve the privacy, noise enabled slicing method is used.