Application of Non-Perturbative SDC Techniques for Microdata

The aim of the present chapter is to consider the problem of producing a safe microdata set by applying global recodings and local suppressions, as discussed in Chapter 1. In our discussion we assume disclosure scenarios of the type discussed in Chapter 2. These scenarios have in common that an intruder is supposed to use a number of low-dimensional combinations of key variables in an attempt to disclose private information. Global recoding and local suppression should be applied in such a way that this type of disclosure is prevented, or at least sufficiently hampered. This can be achieved by making sure unsafe combinations, i.e. with a frequencies below certain assumed threshold values, do not occur. This is precisely the case when the global recodings or local suppressions have yielded a safe microdata set (assuming the disclosure scenario adopted does apply) and with minimum information loss (using an information loss measure as discussed in Chapter 3). Clearly to obtain such a safe microdata through modification of the original unsafe microdata set requires an optimization problem to be solved.