Introduction

Now that you have chosen attribute types you have to choose what action to take with these attributes (columns) in order to achieve required anonymity. In Privacy-Preserving Data Publishing explicit identifier is to be removed because leaving it in data release will be a huge threat to privacy. Columns in QID are usually generalized and columns that contain sensitive attribute are kept as is. Also note that Non-sensitive attributes can be kept as they are since revealing them will not be threat to privacy.

Getting started

In this view you need to specify column actions for each attribute. You can get hints from Introduction above on what these should be. To complete this step, follow these steps:

1. For each attribute (column) choose action that should be taken to anonymize this attribute from combobox. Action should be one of the following:
   • Remove column - remove entire column from the table.
   • Keep as is - does nothing to values. In anonymized table, all values from this column will be exactly as it was in original table.
   • Generalize - generalizes values in this column to more general value.
   • Add noise - Adds random noise to attribute between -ε% and ε% of it's value. This option is only available in ε-Differential privacy and it is intended to be used for sensitive attribute.
2. When you are done, click next to proceed.