LAUSR

Abstract Privacy Preserving Data Publishing (PPDP) is an important aspect of real world scenarios. PPDP moves the researcher in the right direction by maintaining privacy and utility trade-off while publishing the data. This paper presents a concept on dynamic data publishing for multiple sensitive attributes by enhancing KC slice model. Our proposed KC i -slice method completes the data publishing process in two phases. First phase assigns the records into buckets based on the sensitiveness of the attributes, which considers different privacy thresholds on various sensitive attributes. It uses a semantic l-diversity approach to assign the records to the buckets to prevent similarity attacks. The privacy thresholds of all the sensitive attribute values in a bucket are verified. It splits the sensitive attributes into multiple sensitive tables according to the correlation among them. The later phase finds the correlation among quasi attributes. It groups the correlated quasi attributes and also concatenates the SIDs of sensitive attribute values with quasi attribute values. Finally it performs random permutations on the published quasi table. The proposed KC i -slice model enhances the utility and reduces the suppression of multiple sensitive attributes when compared to KC-slice approach.