K- Dataset De-identification Method and Apparatus Using K-anonymity Model

摘要

The present invention relates to a dataset deidentification method using a K-anonymity model and a device thereof. According to the present invention, the dataset deidentification method comprises: a step of confirming attribute value distribution for each column corresponding to a quasi-identifier attribute with respect to a dataset to be deidentified; a first processing step of considering the attribute value distribution for each column to deidentify each column of the dataset to be deidentified; a step of grouping a record with the identical attribute value into a record group when the number of records with the identical attribute value of one or more columns is K or more in the first-processed dataset to be deidentified; and a step of considering attribute value distribution of the column with a non-identical attribute value in the record group to deidentify the column with the non-identical attribute value so as to satisfy K-anonymity requirements. Accordingly, the present invention provides advantages capable of minimizing data crush while the K-anonymity requirements are satisfied and increasing a deidentification speed.