Model of Identification of a Person in Databases of Various Sizes

Depersonalization is a special way of processing personal data, as a result of which it is impossible to identify an individual without additional information. The paper shows that various attributes of an individual are either crucial for their identification (identifiers), or do not matter at all (other data). The problem with any depersonalization method is the choice of attributes-identifiers that will be the objects of influence of this method. Other attributes are not changed. The greater the role of an attribute in identifying an individual, the fewer the number of records with the specified value of this attribute. In this case, the inverse of this number is defined as the probability of identifying an individual by the selected attribute value for a database of a certain volume. To determine the probability of identification over the entire range of values of any attribute, the authors developed an identification model. In addition, the task of the model is to determine the probability of identification for different volumes of the database. The results of calculations allow us to assess the necessity and sufficiency of the depersonalization procedure for specific attributes and volumes of personal data databases.