Package dk.alexandra.fresco.stat.anonymisation

Class LeakyKAnonymity

Object

LeakyKAnonymity

All Implemented Interfaces:

Computation<MultiDimensionalArray<List<DRes<SInt>>>,ProtocolBuilderNumeric>

public class LeakyKAnonymity
extends Object
implements Computation<MultiDimensionalArray<List<DRes<SInt>>>,ProtocolBuilderNumeric>

Compute a k-anonymous version of a dataset. Note that the output leaks the corresponding indices of the individuals in each bucket. To avoid this, the data could be shuffled obliviously before running the computation.

Constructor Summary

Constructors

Constructor	Description
LeakyKAnonymity(Matrix<DRes<SInt>> data, List<DRes<SInt>> sensitive, List<List<DRes<SInt>>> buckets, int k)	Each row in the data set contains the quasi-identifiers of an individual with a corresponding entry in the list of values of the sensitive attribute.

Method Summary

Modifier and Type	Method	Description
DRes<MultiDimensionalArray<List<DRes<SInt>>>>	buildComputation(ProtocolBuilderNumeric builder)

Methods inherited from class Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

LeakyKAnonymity

public LeakyKAnonymity(Matrix<DRes<SInt>> data, List<DRes<SInt>> sensitive, List<List<DRes<SInt>>> buckets, int k)

Each row in the data set contains the quasi-identifiers of an individual with a corresponding entry in the list of values of the sensitive attribute. The buckets indicates the desired generalization of the quasi-identifiers as the upper limits in the corresponding histogram. K is the smallest allowed number of individuals in each bucket.

The output is a histogram on the given buckets with the value in the histogram being a list of size data.getHeight() with a non-zero entry x at index i indicating that the data point at row i is in this bucket and that the corresponding sensitive attribute was x.

Note that the output leaks the corresponding indices of the individuals in each bucket. To avoid this the data could be shuffled obliviously before running the computation.

Method Details

buildComputation

public DRes<MultiDimensionalArray<List<DRes<SInt>>>> buildComputation(ProtocolBuilderNumeric builder)

Specified by:

buildComputation in interface Computation<MultiDimensionalArray<List<DRes<SInt>>>,ProtocolBuilderNumeric>