/** Performs a test. */
@Test
public void testMetric() {
try {
final ARXConfiguration config = ARXConfiguration.create();
config.setMetric(null);
} catch (final NullPointerException e) {
return;
}
Assert.fail();
}
/** Returns a string representing the privacy model */
private String getPrivacyModel(ARXConfiguration config) {
StringBuilder result = new StringBuilder();
result.append("{");
int num = config.getCriteria().size();
int count = 0;
for (PrivacyCriterion c : config.getCriteria()) {
result.append(c.toString());
if (++count < num) {
result.append(", ");
}
}
result.append("}");
return result.toString();
}
/**
* Performs a test.
*
* @throws IOException
*/
@Test
public void testNullHierarchy() throws IOException {
try {
final ARXAnonymizer anonymizer = new ARXAnonymizer();
final Data data = provider.getData();
data.getDefinition().setAttributeType("age", (AttributeType) null);
final ARXConfiguration config = ARXConfiguration.create();
config.addCriterion(new KAnonymity(2));
config.setMaxOutliers(1.2d);
anonymizer.anonymize(data, config);
} catch (final NullPointerException e) {
return;
}
Assert.fail();
}
/**
* Performs a test.
*
* @throws IOException
*/
@Test
public void testEmptyDefinition() throws IOException {
final ARXAnonymizer anonymizer = new ARXAnonymizer();
try {
final ARXConfiguration config = ARXConfiguration.create();
config.addCriterion(new KAnonymity(2));
config.setMaxOutliers(1.2d);
anonymizer.anonymize(provider.getData(), config);
} catch (final IllegalArgumentException e) {
return;
}
Assert.fail();
}
/**
* Performs a test.
*
* @throws IOException
*/
@Test
public void testEmptyDatasetWithAttributeDefinition() throws IOException {
try {
final ARXAnonymizer anonymizer = new ARXAnonymizer();
final Data data = Data.create();
data.getDefinition().setAttributeType("age", AttributeType.IDENTIFYING_ATTRIBUTE);
final ARXConfiguration config = ARXConfiguration.create();
config.addCriterion(new KAnonymity(2));
config.setMaxOutliers(1.2d);
anonymizer.anonymize(provider.getData(), config);
} catch (final IllegalArgumentException e) {
return;
}
Assert.fail();
}
/**
* Performs a test.
*
* @throws IOException
*/
@Test
public void testMissingHierarchyValue() throws IOException {
provider.createDataDefinitionMissing();
final Data data = provider.getData();
final ARXAnonymizer anonymizer = new ARXAnonymizer();
try {
final ARXConfiguration config = ARXConfiguration.create();
config.addCriterion(new KAnonymity(2));
config.setMaxOutliers(0d);
anonymizer.anonymize(data, config);
} catch (final IllegalArgumentException e) {
return;
}
Assert.fail();
}
/**
* Performs a test.
*
* @throws IOException
*/
@Test
public void testEmptyDatasetWithoutAttributeDefinition() throws IOException {
try {
final ARXAnonymizer anonymizer = new ARXAnonymizer();
final Data data = Data.create();
final ARXConfiguration config = ARXConfiguration.create();
config.addCriterion(new KAnonymity(2));
config.setMaxOutliers(1.2d);
anonymizer.anonymize(data, config);
} catch (final IllegalArgumentException e) {
return;
}
Assert.fail();
}
/**
* Creates a new NodeChecker instance.
*
* @param manager The manager
* @param metric The metric
* @param config The anonymization configuration
* @param historyMaxSize The history max size
* @param snapshotSizeDataset The history threshold
* @param snapshotSizeSnapshot The history threshold replacement
*/
public NodeChecker(
final DataManager manager,
final Metric<?> metric,
final ARXConfiguration config,
final int historyMaxSize,
final double snapshotSizeDataset,
final double snapshotSizeSnapshot) {
// Initialize all operators
this.metric = metric;
this.config = config;
data = manager.getDataQI();
final int initialSize = (int) (manager.getDataQI().getDataLength() * 0.01d);
final IntArrayDictionary dictionarySensValue;
final IntArrayDictionary dictionarySensFreq;
if ((config.getRequirements() & ARXConfiguration.REQUIREMENT_DISTRIBUTION) != 0) {
dictionarySensValue = new IntArrayDictionary(initialSize);
dictionarySensFreq = new IntArrayDictionary(initialSize);
} else {
// Just to allow bytecode instrumentation
dictionarySensValue = new IntArrayDictionary(0);
dictionarySensFreq = new IntArrayDictionary(0);
}
history =
new History(
manager.getDataQI().getArray().length,
historyMaxSize,
snapshotSizeDataset,
snapshotSizeSnapshot,
config,
dictionarySensValue,
dictionarySensFreq);
stateMachine = new StateMachine(history);
currentGroupify = new HashGroupify(initialSize, config);
lastGroupify = new HashGroupify(initialSize, config);
transformer =
new Transformer(
manager.getDataQI().getArray(),
manager.getHierarchies(),
manager.getDataSE().getArray(),
config,
dictionarySensValue,
dictionarySensFreq);
}
@Override
public Data transformAndMarkOutliers(final Node node) {
// Apply transition and groupify
currentGroupify.clear();
currentGroupify = transformer.apply(0L, node.getTransformation(), currentGroupify);
// Determine outliers and set infoloss
node.setAnonymous(currentGroupify.isAnonymous());
if (!node.isChecked()) {
node.setChecked();
metric.evaluate(node, currentGroupify);
node.setTagged();
}
// Find outliers
if (config.getAbsoluteMaxOutliers() != 0) {
currentGroupify.markOutliers(transformer.getBuffer());
}
// Return the buffer
return getBuffer();
}
/**
* Entry point.
*
* @param args the arguments
*/
public static void main(String[] args) throws IOException {
// Define data
Data data = getData();
// Define attribute types
data.getDefinition().setAttributeType("age", getHierarchyAge());
data.getDefinition().setAttributeType("zipcode", getHierarchyZipcode());
data.getDefinition().setAttributeType("disease1", AttributeType.SENSITIVE_ATTRIBUTE);
data.getDefinition().setAttributeType("disease2", AttributeType.SENSITIVE_ATTRIBUTE);
// Create an instance of the anonymizer
ARXAnonymizer anonymizer = new ARXAnonymizer();
ARXConfiguration config = ARXConfiguration.create();
config.addPrivacyModel(new KAnonymity(3));
config.addPrivacyModel(
new HierarchicalDistanceTCloseness("disease1", 0.6d, getHierarchyDisease()));
config.addPrivacyModel(new RecursiveCLDiversity("disease2", 3d, 2));
config.setMaxOutliers(0d);
config.setQualityModel(Metric.createEntropyMetric());
// Now anonymize
ARXResult result = anonymizer.anonymize(data, config);
// Print info
printResult(result, data);
// Process results
if (result.getGlobalOptimum() != null) {
System.out.println(" - Transformed data:");
Iterator<String[]> transformed = result.getOutput(false).iterator();
while (transformed.hasNext()) {
System.out.print(" ");
System.out.println(Arrays.toString(transformed.next()));
}
}
}
/**
* Returns the test cases.
*
* @return
*/
@Parameters(name = "{index}:[{0}]")
public static Collection<Object[]> cases() {
return Arrays.asList(
new Object[][] {
/* 0 */
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(
new EntropyLDiversity("occupation", 5, EntropyEstimator.GRASSBERGER)),
"occupation",
"./data/adult.csv",
216092.124036387,
new int[] {1, 0, 1, 0, 3, 2, 2, 0},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(
new EntropyLDiversity("occupation", 100, EntropyEstimator.SHANNON)),
"occupation",
"./data/adult.csv",
0.0d,
null,
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(
new EntropyLDiversity("occupation", 5, EntropyEstimator.GRASSBERGER)),
"occupation",
"./data/adult.csv",
324620.5269918692,
new int[] {1, 1, 1, 1, 3, 2, 2, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.05d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(
new EntropyLDiversity("occupation", 3, EntropyEstimator.GRASSBERGER)),
"occupation",
"./data/adult.csv",
180347.4325366015,
new int[] {0, 0, 1, 1, 2, 2, 2, 0},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(
new EntropyLDiversity("occupation", 5, EntropyEstimator.SHANNON)),
"occupation",
"./data/adult.csv",
228878.2039109517,
new int[] {1, 0, 1, 1, 2, 2, 2, 1},
true)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.1d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(
new EntropyLDiversity("occupation", 100, EntropyEstimator.GRASSBERGER)),
"occupation",
"./data/adult.csv",
0.0d,
null,
true)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createDiscernabilityMetric(true))
.addPrivacyModel(
new EntropyLDiversity("RAMNTALL", 5, EntropyEstimator.GRASSBERGER)),
"RAMNTALL",
"./data/cup.csv",
1833435.0,
new int[] {4, 0, 1, 0, 1, 3, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.03d, Metric.createDiscernabilityMetric(true))
.addPrivacyModel(
new EntropyLDiversity("RAMNTALL", 100, EntropyEstimator.GRASSBERGER)),
"RAMNTALL",
"./data/cup.csv",
4.5168281E7,
new int[] {4, 4, 0, 0, 1, 3, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createDiscernabilityMetric(true))
.addPrivacyModel(new EntropyLDiversity("RAMNTALL", 5)),
"RAMNTALL",
"./data/cup.csv",
3.01506905E8,
new int[] {4, 4, 1, 1, 1, 4, 4},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createDiscernabilityMetric(true))
.addPrivacyModel(new EntropyLDiversity("RAMNTALL", 3)),
"RAMNTALL",
"./data/cup.csv",
9.2264547E7,
new int[] {4, 4, 1, 0, 1, 4, 4},
false)
},
/* 10 */ {
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createDiscernabilityMetric(true))
.addPrivacyModel(
new EntropyLDiversity("RAMNTALL", 5, EntropyEstimator.SHANNON)),
"RAMNTALL",
"./data/cup.csv",
2823649.0,
new int[] {4, 0, 0, 1, 1, 3, 1},
true)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.1d, Metric.createDiscernabilityMetric(true))
.addPrivacyModel(
new EntropyLDiversity("RAMNTALL", 100, EntropyEstimator.GRASSBERGER)),
"RAMNTALL",
"./data/cup.csv",
3.4459973E7,
new int[] {5, 0, 0, 2, 1, 2, 1},
true)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(
new EntropyLDiversity("EDUC", 5, EntropyEstimator.GRASSBERGER)),
"EDUC",
"./data/ihis.csv",
7735322.29514608,
new int[] {0, 0, 0, 1, 3, 0, 0, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(
new EntropyLDiversity("EDUC", 2, EntropyEstimator.GRASSBERGER)),
"EDUC",
"./data/ihis.csv",
5428093.534997522,
new int[] {0, 0, 0, 0, 2, 0, 0, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(new EntropyLDiversity("EDUC", 5, EntropyEstimator.SHANNON)),
"EDUC",
"./data/ihis.csv",
1.2258628558792587E7,
new int[] {0, 0, 0, 3, 3, 2, 0, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(
new EntropyLDiversity("EDUC", 100, EntropyEstimator.GRASSBERGER)),
"EDUC",
"./data/ihis.csv",
0.0d,
null,
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(
new EntropyLDiversity("EDUC", 5, EntropyEstimator.GRASSBERGER)),
"EDUC",
"./data/ihis.csv",
7735322.29514608,
new int[] {0, 0, 0, 1, 3, 0, 0, 1},
true)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.02d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(new EntropyLDiversity("EDUC", 3, EntropyEstimator.SHANNON)),
"EDUC",
"./data/ihis.csv",
7578152.206004559,
new int[] {0, 0, 0, 2, 2, 0, 0, 1},
true)
},
});
}
/**
* Return prosecutor risk threshold, 1 if there is none
*
* @return
*/
public double getEstimatedProsecutorRisk() {
return Math.min(
1.0d / (double) getHistogram().getHistogram()[0],
config != null && anonymous ? config.getRiskThresholdProsecutor() : 1d);
}
/**
* Return marketer risk threshold, 1 if there is none
*
* @return
*/
public double getEstimatedMarketerRisk() {
return Math.min(
1.0d / getHistogram().getAvgClassSize(),
config != null && anonymous ? config.getRiskThresholdMarketer() : 1d);
}
/**
* Entry point.
*
* @param args the arguments
*/
public static void main(String[] args) throws IOException {
// Define data
DefaultData data = Data.create();
data.add("age", "gender", "zipcode");
data.add("45", "female", "81675");
data.add("34", "male", "81667");
data.add("66", "male", "81925");
data.add("70", "female", "81931");
data.add("34", "female", "81931");
data.add("70", "male", "81931");
data.add("45", "male", "81931");
// Define hierarchies
DefaultHierarchy age = Hierarchy.create();
age.add("34", "<50", "*");
age.add("45", "<50", "*");
age.add("66", ">=50", "*");
age.add("70", ">=50", "*");
DefaultHierarchy gender = Hierarchy.create();
gender.add("male", "*");
gender.add("female", "*");
// Only excerpts for readability
DefaultHierarchy zipcode = Hierarchy.create();
zipcode.add("81667", "8166*", "816**", "81***", "8****", "*****");
zipcode.add("81675", "8167*", "816**", "81***", "8****", "*****");
zipcode.add("81925", "8192*", "819**", "81***", "8****", "*****");
zipcode.add("81931", "8193*", "819**", "81***", "8****", "*****");
data.getDefinition().setAttributeType("age", age);
data.getDefinition().setAttributeType("gender", gender);
data.getDefinition().setAttributeType("zipcode", zipcode);
// Create an instance of the anonymizer
ARXAnonymizer anonymizer = new ARXAnonymizer();
ARXConfiguration config = ARXConfiguration.create();
config.addCriterion(new KAnonymity(2));
config.setMaxOutliers(0d);
ARXResult result = anonymizer.anonymize(data, config);
// Print info
printResult(result, data);
// Print input
System.out.println(" - Input data:");
Iterator<String[]> original = data.getHandle().iterator();
while (original.hasNext()) {
System.out.print(" ");
System.out.println(Arrays.toString(original.next()));
}
// Print results
System.out.println(" - Transformed data:");
Iterator<String[]> transformed = result.getOutput(false).iterator();
while (transformed.hasNext()) {
System.out.print(" ");
System.out.println(Arrays.toString(transformed.next()));
}
// Print frequencies
StatisticsFrequencyDistribution distribution;
System.out.println(" - Distribution of attribute 'age' in input:");
distribution = data.getHandle().getStatistics().getFrequencyDistribution(0, false);
System.out.println(" " + Arrays.toString(distribution.values));
System.out.println(" " + Arrays.toString(distribution.frequency));
// Print frequencies
System.out.println(" - Distribution of attribute 'age' in output:");
distribution = result.getOutput(false).getStatistics().getFrequencyDistribution(0, true);
System.out.println(" " + Arrays.toString(distribution.values));
System.out.println(" " + Arrays.toString(distribution.frequency));
// Print contingency tables
StatisticsContingencyTable contingency;
System.out.println(" - Contingency of attribute 'gender' and 'zipcode' in input:");
contingency = data.getHandle().getStatistics().getContingencyTable(0, true, 2, true);
System.out.println(" " + Arrays.toString(contingency.values1));
System.out.println(" " + Arrays.toString(contingency.values2));
while (contingency.iterator.hasNext()) {
Entry e = contingency.iterator.next();
System.out.println(" [" + e.value1 + ", " + e.value2 + ", " + e.frequency + "]");
}
// Print contingency tables
System.out.println(" - Contingency of attribute 'gender' and 'zipcode' in output:");
contingency = result.getOutput(false).getStatistics().getContingencyTable(0, true, 2, true);
System.out.println(" " + Arrays.toString(contingency.values1));
System.out.println(" " + Arrays.toString(contingency.values2));
while (contingency.iterator.hasNext()) {
Entry e = contingency.iterator.next();
System.out.println(" [" + e.value1 + ", " + e.value2 + ", " + e.frequency + "]");
}
}
/**
* Returns the test cases.
*
* @return
*/
@Parameters(name = "{index}:[{0}]")
public static Collection<Object[]> cases() {
return Arrays.asList(
new Object[][] {
/* 0 */ {
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(new KAnonymity(5)),
"./data/adult.csv",
255559.85455731067,
new int[] {1, 0, 1, 1, 3, 2, 2, 0, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(new KAnonymity(100)),
"./data/adult.csv",
379417.3460570988,
new int[] {1, 1, 1, 1, 3, 2, 2, 1, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(new KAnonymity(5)),
"./data/adult.csv",
407289.5388925293,
new int[] {1, 2, 1, 1, 3, 2, 2, 1, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(new KAnonymity(100)),
"./data/adult.csv",
453196.8932458743,
new int[] {0, 4, 1, 1, 3, 2, 2, 1, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(new KAnonymity(5)),
"./data/adult.csv",
255559.85455731067,
new int[] {1, 0, 1, 1, 3, 2, 2, 0, 1},
true)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(new KAnonymity(100)),
"./data/adult.csv",
379417.3460570988,
new int[] {1, 1, 1, 1, 3, 2, 2, 1, 1},
true)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(new KAnonymity(5)),
"./data/cup.csv",
1764006.4033760305,
new int[] {2, 4, 0, 1, 0, 4, 4, 4},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(new KAnonymity(100)),
"./data/cup.csv",
1994002.8308631124,
new int[] {3, 4, 1, 1, 0, 4, 4, 4},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(new KAnonymity(5)),
"./data/cup.csv",
2445878.424834677,
new int[] {4, 4, 1, 1, 1, 4, 4, 4},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(new KAnonymity(100)),
"./data/cup.csv",
2517471.5816586106,
new int[] {5, 4, 1, 0, 1, 4, 4, 4},
false)
},
/* 10 */ {
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(new KAnonymity(5)),
"./data/cup.csv",
1764006.4033760305,
new int[] {2, 4, 0, 1, 0, 4, 4, 4},
true)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(new KAnonymity(100)),
"./data/cup.csv",
2001343.4737485605,
new int[] {3, 4, 1, 1, 0, 1, 2, 1},
true)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createDiscernabilityMetric(true))
.addPrivacyModel(new KAnonymity(5)),
"./data/fars.csv",
4469271.0,
new int[] {0, 2, 2, 2, 1, 2, 1, 0},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createDiscernabilityMetric(true))
.addPrivacyModel(new KAnonymity(100)),
"./data/fars.csv",
5.6052481E7,
new int[] {0, 2, 3, 3, 1, 2, 2, 2},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createDiscernabilityMetric(true))
.addPrivacyModel(new KAnonymity(5)),
"./data/fars.csv",
1.42377891E8,
new int[] {1, 2, 3, 3, 1, 2, 1, 2},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createDiscernabilityMetric(true))
.addPrivacyModel(new KAnonymity(100)),
"./data/fars.csv",
4.36925397E8,
new int[] {5, 2, 3, 3, 1, 2, 0, 2},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createDiscernabilityMetric(true))
.addPrivacyModel(new KAnonymity(5)),
"./data/fars.csv",
4469271.0,
new int[] {0, 2, 2, 2, 1, 2, 1, 0},
true)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.04d, Metric.createDiscernabilityMetric(true))
.addPrivacyModel(new KAnonymity(100)),
"./data/fars.csv",
5.6052481E7,
new int[] {0, 2, 3, 3, 1, 2, 2, 2},
true)
},
});
}
@Override
protected void initializeInternal(
final DataDefinition definition,
final Data input,
final GeneralizationHierarchy[] ahierarchies,
final ARXConfiguration config) {
// Obtain dictionary
final Dictionary dictionary = input.getDictionary();
// Obtain research subset
RowSet rSubset = null;
if (config.containsCriterion(DPresence.class)) {
Set<DPresence> crits = config.getCriteria(DPresence.class);
if (crits.size() > 1) {
throw new IllegalArgumentException("Only one d-presence criterion supported!");
}
for (DPresence dPresence : crits) {
rSubset = dPresence.getSubset().getSet();
}
}
// Create reference to the hierarchies
final int[][] data = input.getArray();
hierarchies = new int[data[0].length][][];
for (int i = 0; i < ahierarchies.length; i++) {
hierarchies[i] = ahierarchies[i].getArray();
// Column -> Id -> Level -> Output
}
// Initialize counts
cardinalities = new int[data[0].length][][];
for (int i = 0; i < cardinalities.length; i++) {
cardinalities[i] =
new int[dictionary.getMapping()[i].length][ahierarchies[i].getArray()[0].length];
// Column -> Id -> Level -> Count
}
for (int i = 0; i < data.length; i++) {
// only use the rows contained in the research subset
if (rSubset == null || rSubset.contains(i)) {
final int[] row = data[i];
for (int column = 0; column < row.length; column++) {
cardinalities[column][row[column]][0]++;
}
}
}
// Create counts for other levels
for (int column = 0; column < hierarchies.length; column++) {
final int[][] hierarchy = hierarchies[column];
for (int in = 0; in < hierarchy.length; in++) {
final int cardinality = cardinalities[column][in][0];
for (int level = 1; level < hierarchy[in].length; level++) {
final int out = hierarchy[in][level];
cardinalities[column][out][level] += cardinality;
}
}
}
// Create a cache for the results
cache = new double[hierarchies.length][];
for (int i = 0; i < cache.length; i++) {
cache[i] = new double[ahierarchies[i].getArray()[0].length];
Arrays.fill(cache[i], NA);
}
}
/**
* Returns the test cases.
*
* @return
* @throws IOException
*/
@Parameters(name = "{index}:[{0}]")
public static Collection<Object[]> cases() throws IOException {
return Arrays.asList(
new Object[][] {
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.05d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(
new KMap(
3,
0.01d,
ARXPopulationModel.create(Region.USA),
CellSizeEstimator.ZERO_TRUNCATED_POISSON)),
"occupation",
"./data/adult.csv",
130804.5332092598,
new int[] {0, 0, 1, 1, 0, 2, 2, 0},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.05d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(
new KMap(
1000,
0.01d,
ARXPopulationModel.create(Region.USA),
CellSizeEstimator.ZERO_TRUNCATED_POISSON)),
"occupation",
"./data/adult.csv",
151894.1394841501,
new int[] {0, 0, 1, 1, 1, 2, 1, 0},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createAECSMetric())
.addPrivacyModel(
new KMap(
5,
DataSubset.create(
Data.create("./data/adult.csv", StandardCharsets.UTF_8, ';'),
Data.create(
"./data/adult_subset.csv", StandardCharsets.UTF_8, ';')))),
"occupation",
"./data/adult.csv",
45.014925373134325,
new int[] {1, 0, 1, 2, 3, 2, 2, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(
new KMap(
3,
DataSubset.create(
Data.create("./data/adult.csv", StandardCharsets.UTF_8, ';'),
Data.create(
"./data/adult_subset.csv", StandardCharsets.UTF_8, ';')))),
"occupation",
"./data/adult.csv",
23387.494246375998,
new int[] {0, 0, 1, 2, 3, 2, 2, 0},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(
new KMap(
5,
DataSubset.create(
Data.create("./data/adult.csv", StandardCharsets.UTF_8, ';'),
Data.create(
"./data/adult_subset.csv", StandardCharsets.UTF_8, ';')))),
"occupation",
"./data/adult.csv",
28551.7222913157,
new int[] {1, 0, 1, 2, 3, 2, 2, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.05d, Metric.createAECSMetric())
.addPrivacyModel(
new KMap(
20,
DataSubset.create(
Data.create("./data/adult.csv", StandardCharsets.UTF_8, ';'),
Data.create(
"./data/adult_subset.csv", StandardCharsets.UTF_8, ';')))),
"occupation",
"./data/adult.csv",
11.424242424242424,
new int[] {1, 0, 1, 1, 3, 2, 1, 0},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.05d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(
new KMap(
7,
DataSubset.create(
Data.create("./data/adult.csv", StandardCharsets.UTF_8, ';'),
Data.create(
"./data/adult_subset.csv", StandardCharsets.UTF_8, ';')))),
"occupation",
"./data/adult.csv",
17075.7181747451,
new int[] {0, 0, 1, 1, 2, 2, 2, 0},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.05d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(
new KMap(
3,
DataSubset.create(
Data.create("./data/adult.csv", StandardCharsets.UTF_8, ';'),
Data.create(
"./data/adult_subset.csv", StandardCharsets.UTF_8, ';')))),
"occupation",
"./data/adult.csv",
15121.633326877098,
new int[] {0, 0, 1, 1, 1, 2, 1, 0},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createAECSMetric())
.addPrivacyModel(
new KMap(
5,
DataSubset.create(
Data.create("./data/adult.csv", StandardCharsets.UTF_8, ';'),
Data.create(
"./data/adult_subset.csv", StandardCharsets.UTF_8, ';')))),
"occupation",
"./data/adult.csv",
45.014925373134325,
new int[] {1, 0, 1, 2, 3, 2, 2, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(
new KMap(
2,
DataSubset.create(
Data.create("./data/adult.csv", StandardCharsets.UTF_8, ';'),
Data.create(
"./data/adult_subset.csv", StandardCharsets.UTF_8, ';')))),
"occupation",
"./data/adult.csv",
23108.1673304724,
new int[] {1, 0, 1, 1, 3, 2, 2, 0},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.0d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(
new KMap(
10,
DataSubset.create(
Data.create("./data/adult.csv", StandardCharsets.UTF_8, ';'),
Data.create(
"./data/adult_subset.csv", StandardCharsets.UTF_8, ';')))),
"occupation",
"./data/adult.csv",
30238.2081484441,
new int[] {0, 1, 1, 2, 3, 2, 2, 0},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.05d, Metric.createAECSMetric())
.addPrivacyModel(
new KMap(
10,
DataSubset.create(
Data.create("./data/adult.csv", StandardCharsets.UTF_8, ';'),
Data.create(
"./data/adult_subset.csv", StandardCharsets.UTF_8, ';')))),
"occupation",
"./data/adult.csv",
7.215311004784689,
new int[] {0, 0, 1, 1, 3, 2, 1, 0},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.05d, Metric.createPrecomputedEntropyMetric(0.1d, true))
.addPrivacyModel(
new KMap(
5,
DataSubset.create(
Data.create("./data/adult.csv", StandardCharsets.UTF_8, ';'),
Data.create(
"./data/adult_subset.csv", StandardCharsets.UTF_8, ';')))),
"occupation",
"./data/adult.csv",
17053.8743069776,
new int[] {0, 0, 1, 0, 2, 2, 2, 1},
false)
},
{
new ARXAnonymizationTestCase(
ARXConfiguration.create(0.05d, Metric.createPrecomputedEntropyMetric(0.1d, false))
.addPrivacyModel(
new KMap(
3,
DataSubset.create(
Data.create("./data/adult.csv", StandardCharsets.UTF_8, ';'),
Data.create(
"./data/adult_subset.csv", StandardCharsets.UTF_8, ';')))),
"occupation",
"./data/adult.csv",
15121.633326877098,
new int[] {0, 0, 1, 1, 1, 2, 1, 0},
false)
},
});
}
/**
* Entry point.
*
* @param args the arguments
* @throws IOException
*/
public static void main(String[] args) throws IOException {
// Define data
DefaultData data = Data.create();
data.add("zipcode", "age", "disease");
data.add("47677", "29", "gastric ulcer");
data.add("47602", "22", "gastritis");
data.add("47678", "27", "stomach cancer");
data.add("47905", "43", "gastritis");
data.add("47909", "52", "flu");
data.add("47906", "47", "bronchitis");
data.add("47605", "30", "bronchitis");
data.add("47673", "36", "pneumonia");
data.add("47607", "32", "stomach cancer");
// Define hierarchies
DefaultHierarchy age = Hierarchy.create();
age.add("29", "<=40", "*");
age.add("22", "<=40", "*");
age.add("27", "<=40", "*");
age.add("43", ">40", "*");
age.add("52", ">40", "*");
age.add("47", ">40", "*");
age.add("30", "<=40", "*");
age.add("36", "<=40", "*");
age.add("32", "<=40", "*");
// Only excerpts for readability
DefaultHierarchy zipcode = Hierarchy.create();
zipcode.add("47677", "4767*", "476**", "47***", "4****", "*****");
zipcode.add("47602", "4760*", "476**", "47***", "4****", "*****");
zipcode.add("47678", "4767*", "476**", "47***", "4****", "*****");
zipcode.add("47905", "4790*", "479**", "47***", "4****", "*****");
zipcode.add("47909", "4790*", "479**", "47***", "4****", "*****");
zipcode.add("47906", "4790*", "479**", "47***", "4****", "*****");
zipcode.add("47605", "4760*", "476**", "47***", "4****", "*****");
zipcode.add("47673", "4767*", "476**", "47***", "4****", "*****");
zipcode.add("47607", "4760*", "476**", "47***", "4****", "*****");
// Define sensitive value hierarchy
DefaultHierarchy disease = Hierarchy.create();
disease.add(
"flu",
"respiratory infection",
"vascular lung disease",
"respiratory & digestive system disease");
disease.add(
"pneumonia",
"respiratory infection",
"vascular lung disease",
"respiratory & digestive system disease");
disease.add(
"bronchitis",
"respiratory infection",
"vascular lung disease",
"respiratory & digestive system disease");
disease.add(
"pulmonary edema",
"vascular lung disease",
"vascular lung disease",
"respiratory & digestive system disease");
disease.add(
"pulmonary embolism",
"vascular lung disease",
"vascular lung disease",
"respiratory & digestive system disease");
disease.add(
"gastric ulcer",
"stomach disease",
"digestive system disease",
"respiratory & digestive system disease");
disease.add(
"stomach cancer",
"stomach disease",
"digestive system disease",
"respiratory & digestive system disease");
disease.add(
"gastritis",
"stomach disease",
"digestive system disease",
"respiratory & digestive system disease");
disease.add(
"colitis",
"colon disease",
"digestive system disease",
"respiratory & digestive system disease");
disease.add(
"colon cancer",
"colon disease",
"digestive system disease",
"respiratory & digestive system disease");
data.getDefinition().setAttributeType("age", age);
data.getDefinition().setAttributeType("zipcode", zipcode);
data.getDefinition().setAttributeType("disease", AttributeType.SENSITIVE_ATTRIBUTE);
// Create an instance of the anonymizer
ARXAnonymizer anonymizer = new ARXAnonymizer();
ARXConfiguration config = ARXConfiguration.create();
config.addCriterion(new KAnonymity(3));
config.addCriterion(new HierarchicalDistanceTCloseness("disease", 0.6d, disease));
config.setMaxOutliers(0d);
config.setMetric(Metric.createEntropyMetric());
// Now anonymize
ARXResult result = anonymizer.anonymize(data, config);
// Print info
printResult(result, data);
// Process results
System.out.println(" - Transformed data:");
Iterator<String[]> transformed = result.getOutput(false).iterator();
while (transformed.hasNext()) {
System.out.print(" ");
System.out.println(Arrays.toString(transformed.next()));
}
}
