Whenever there is some ­sensational data released by government or corporates often we see lot of controversy happening around. Some people complain of invading their privacy and demand to remove offensive content from the data declaration. Imagine the trauma a HIV patient goes through when it is revealed in front of the world that he is infected or more recently the negative effect of social media where anyone from anywhere can post offensive and provocative content. All such uncontrolled data publishing should be screened and monitored so that privacy of an individual or an institution is preserved.  Privacy preserving data publishing ensures that privacy is preserved while the data gets revealed or shared; it is done using some techniques and methods according to corresponding situation and scenarios. 

Approaches in PPDP: 

Micro Data Publishing:  Includes public census for government records, the details such as name, DOB, job profiles etc. are taken from members of a family. The disclosure of most of this data would be minimal. 

Data Anonymization: There is a huge amount of risks associated with the disclosure of sensitive data, it must be anonymized before publishing. All explicit and quasi identifiers are replaced with mellowed down and inconsistent data. 

The data publishing process includes various persons such as 

  • The individual from whom data is collected
  • The collection agent
  • The adversary
  • The end user who uses the data 

The components taken care while publishing the data are sanitization mechanism, privacy criterion and utility metric. Disclosures can be both positive and negative and each is denoted by separate anonymization algorithms. Then there is also uninformative principle in which additional info are added by the adversary. PDPP also includes algorithms and analysis methods like k-Anonymity, I-Diversity, Closeness, utility analysis, anonymization techniques, distance measures and other experiments. Each technique adapts different methods to preserve data privacy.

Download Privacy Preserving Data mining (PDDP) Seminar Report .