Securing the Privacy
We all need privacy for certain information that has to be send. Do you know how to attain this privacy for publishing the data? Privacy is a very important issue when one desires to make use of information that involves individual’s sensitive information. Research on securing the privacy of people and the confidentiality of information has received contributions from several fields, including statistics, social science, economics, and computer science.
Looking at the privacy criteria which provides the required safety guarantees, provide an algorithm that sanitize information to make it safe for release while protecting useful information, and also discuss the methods of analyzing the sanitized data. Sometimes data in its original form may contain sensitive data about people, and publishing such data may violate the privacy of the individuals. Privacy-preserving data publishing (PPDP) gives ways and tools for publishing useful information which also preserve privacy of the data.
In the most simple type of PPDP, the data publisher contains a table of the form D, where specific identifier may be a set of attributes, like name and social security number (SSN), containing data that explicitly identifies record owners; quasi identifier (QID) may be a set of attributes that might potentially determine record owners; Sensitive Attributes consists of sensitive person specific data like illness, salary, and disability status; and Non-Sensitive Attributes contains all attributes that don’t fall into the previous 3 classes. Also the four sets of attributes are disjoint.
Most works assume that every record in the table represents a definite owner of a record. Anonymization refers to the PPDP approach that seeks to cover the identity and/or the sensitive information of record owners. Assuming that sensitive information should be preserved for information analysis.
Privacy-Preserving Data Publishing
Privacy protection provided a very stringent definition:. access to the published information shouldn’t modify the attacker to learn anything more regarding any target victim compared to no access to the database. Even with the presence of attacker’s information obtained from other sources. The first class considers that a privacy threat happens when an attacker is able to link a record owner to a record in a published data table, to a sensitive data in a published information table, or to the published information table itself.
We usually consider the privacy models in two categories depending on the attacking principles. Now let us have a look at these.
The first class considers that a privacy threat happens when an adversary is able to link a record owner to a record in a published information table, to a sensitive attribute in a published information table, or to the published information table itself. We tend to call these respectively as, record, attribute, and table linkage. In all 3 kinds of linkages, we assume that the individual is aware of the QID of the victim.
Record and Attribute Linkages
In record and attribute linkages, we also assume that the adversary is aware of the victim’s record is in the released table. And seeks to spot the victim’s record and/or sensitive information from the table. In table linkage, the attack seeks to identify the presence or absence of the victim’s record in the released table. The data table is considered as privacy-preserving if it will effectively prevent the adversary from successfully performing these linkages.
The second class aims at achieving the uninformative principle. The published table should provide the person with little extra data beyond the information. Several privacy models in this family don’t explicitly classify attributes in a information table into QID and Sensitive Attributes. But some of them may also thwart the sensitive linkages in the initial class, so the 2 classes overlap.0