An Effective Algorithm for Multiple Sensitive Attributes to Preserve Data Privacy

To make various statistics about any medical data huge amount of data gets collected from the public. The collected information contains sensitive information about the individual. Individual's sensitive information includes data such as previous health condition, travel history, political affinity and family background. The collected data is published to perform various statistics, carry out research and to make decisions. However, there might be an intruder in the crowd who uses the published data to gain more information about a particular person that leads to privacy threat. Privacy Preserving Data Publishing (PPDP) aims at anonymizing such data so that it becomes difficult for an intruder to infer a various attribute of particular person from the published data and to gain more information about a him. When the data is anonymized, there are chances of data loss, hence the anonymization algorithm, must balance between the anonymity and the utility of the published data. Most of the existing privacy algorithm concentrate on preserving privacy of single sensitive attributes, but many domains require anonymizing multiple sensitive attributes such as healthcare/ hospital dataset, census dataset etc., that need to be preserved. In this paper an effective algorithm is proposed that preserves the privacy of multiple sensitive attributes with minimal information loss.