摘要 |
<P>PROBLEM TO BE SOLVED: To quickly derive an optimum table as much as possible by reducing processing loads and unnecessary processing. <P>SOLUTION: Inputted data are classified to quasi-identifiers or important information by each attribute, a table composed of all quasi-identifiers and a table of all important information are created from the inputted data, priority (weighting) of respective data is set in each attribute of the data while considering request conditions of users utilizing public information, and evaluation points of respective data are calculated based on the set priority (weighting). Data are generalized so as to satisfy k-anonymity by a simple attribute in each attribute, the generalized data are processed by top-down processing, and then the data are processed by bottom-up processing. The top-down processing and the bottom-up processing are repeated, and after determining whether or not the table composed of all quasi-identifiers satisfies k-anonymity and determining whether or not the table composed of all important information satisfies l-diversity, an optimum data set is outputted. <P>COPYRIGHT: (C)2012,JPO&INPIT |