anonymizer
Class QIDAttribute

java.lang.Object
  anonymizer.QIDAttribute

public class QIDAttribute
extends java.lang.Object

Data structure that holds all relevant information related to a quasi-identifier attribute. Namely, attribute index in the input, the value generalization hierarchy (read from the configuration file), string->integer domain mapping for categorical attributes.

Field Summary

java.util.Hashtable<java.lang.String,java.lang.Integer>	catDomMapping Mapping from the categorical domain to discrete-valued numerical (integer) domain for categorical attributes
private java.util.Hashtable<java.lang.String,java.lang.String[]>	childLookup Lookup to find children (specializations) of an interval
private java.util.Hashtable<java.lang.String,java.lang.String[]>	generalizationSeq Maps each VGH entry to a list of ground-domain values (i.e., not generalized values)
int	index Index of the attribute
private java.util.LinkedList<Interval>	leafIntervals List of leaf intervals in the VGH
private java.util.Hashtable<java.lang.String,java.lang.Integer>	nonLeaves Maps every non-leaf VGH node to a unique index, assigned by a breadth-first traversal of the VGH
private java.util.Hashtable<java.lang.String,java.lang.String>	parentLookup Lookup to find the generalization of an interval
private java.lang.String	suppValue Suppression value for the VGH

Constructor Summary

QIDAttribute(int index, org.w3c.dom.Node att)
Class constructor

Method Summary

java.lang.String	generalize(java.lang.String value) Retrieves the generalization of any QID value (leaf or non-leaf)
boolean	generalizesTo(java.lang.String value, java.lang.String generalization) Checks whether a generalization matches a value
private int	getDepth(java.lang.String leafInterval) Calculates the depth of a leaf interval (i.e., length of the path to the suppression value)
java.lang.String[]	getGeneralizationSequence(java.lang.String val) Get the list of ground-domain values that the provided value represents (applies only to categorical attributes mapped to a discrete domain)
int[]	getLeafCategories(java.lang.String val)
java.util.Hashtable<java.lang.String,java.lang.Integer>	getNonLeaves() Get the non-leaf nodes of the VGH
java.lang.String	getSup() Getter for supression value
int	getVGHDepth(boolean validateDGH) Calculates the depth of the value generalization hierarchy
boolean	isDomainGeneralizationHierarchy() Checks whether the VGH for this qid-attribute corresponds to a domain generalization hierarchy.
private void	parseMapping(org.w3c.dom.Node map) Assign the mapping from the categorical domain to discrete valued numerical domain
private void	parseVGH(org.w3c.dom.Node vgh) Obtain the value-generalization hierarchy for the attribute
java.lang.String[]	specialize(java.lang.String value) Retrieves higher granularity values of some generalized value

Methods inherited from class java.lang.Object

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

Field Detail

catDomMapping

public java.util.Hashtable<java.lang.String,java.lang.Integer> catDomMapping

Mapping from the categorical domain to discrete-valued numerical (integer) domain for categorical attributes

index

public int index

Index of the attribute

suppValue

private java.lang.String suppValue

Suppression value for the VGH

parentLookup

private java.util.Hashtable<java.lang.String,java.lang.String> parentLookup

Lookup to find the generalization of an interval

childLookup

private java.util.Hashtable<java.lang.String,java.lang.String[]> childLookup

Lookup to find children (specializations) of an interval

leafIntervals

private java.util.LinkedList<Interval> leafIntervals

List of leaf intervals in the VGH

nonLeaves

private java.util.Hashtable<java.lang.String,java.lang.Integer> nonLeaves

Maps every non-leaf VGH node to a unique index, assigned by a breadth-first traversal of the VGH

generalizationSeq

private java.util.Hashtable<java.lang.String,java.lang.String[]> generalizationSeq

Maps each VGH entry to a list of ground-domain values (i.e., not generalized values)

Constructor Detail

QIDAttribute

public QIDAttribute(int index,
                    org.w3c.dom.Node att)
             throws java.lang.Exception

Class constructor

Parameters:

att - The node that describes a QID attribute in the configuration file

index - Index of the attribute

Throws:

java.lang.Exception

Method Detail

parseMapping

private void parseMapping(org.w3c.dom.Node map)
                   throws java.lang.Exception

Assign the mapping from the categorical domain to discrete valued numerical domain

Parameters:

map - Root node that contains (category, integer-value) pairs

Throws:

java.lang.Exception

parseVGH

private void parseVGH(org.w3c.dom.Node vgh)
               throws java.lang.Exception

Obtain the value-generalization hierarchy for the attribute

Parameters:

vgh - Root node that contains the suppression value and all its chilren (specializations)

Throws:

java.lang.Exception

isDomainGeneralizationHierarchy

public boolean isDomainGeneralizationHierarchy()

Checks whether the VGH for this qid-attribute corresponds to a domain generalization hierarchy. This basically requires that all leave values in the VGH be at the same depth from the root (i.e., the suppression value). The check should be carried out all quasi-identifier attributes of all full-domain anonymization methods (e.g., Datafly, Incognito, etc.).

Returns:

true if VGH corresponds to a DGH, false otherwise

getVGHDepth

public int getVGHDepth(boolean validateDGH)
                throws java.lang.Exception

Calculates the depth of the value generalization hierarchy

Parameters:

validateDGH - if true, this method first checks whether the domain generalization hiearchy represents a value generalization hierarchy (i.e., whether all leaves are at the same depth or not)

Returns:

Depth of the VGH

Throws:

java.lang.Exception - If input validation fails

getDepth

private int getDepth(java.lang.String leafInterval)

Calculates the depth of a leaf interval (i.e., length of the path to the suppression value)

Parameters:

leafInterval - String representation of a leaf interval

Returns:

depth of the leaf interval

getSup

public java.lang.String getSup()

Getter for supression value

Returns:

root value of the DGH

generalize

public java.lang.String generalize(java.lang.String value)

Retrieves the generalization of any QID value (leaf or non-leaf)

Parameters:

value - some value from the QID's domain

Returns:

immediate parent of the value (or itself if DGH root is specified)

specialize

public java.lang.String[] specialize(java.lang.String value)
                              throws java.lang.Exception

Retrieves higher granularity values of some generalized value

Parameters:

value - some non-leaf value from DGH

Returns:

array of possible specializations of the value, or null if the attribute is continuous and the value is not a node of the VGH

Throws:

java.lang.Exception

generalizesTo

public boolean generalizesTo(java.lang.String value,
                             java.lang.String generalization)
                      throws java.lang.Exception

Checks whether a generalization matches a value

Parameters:

value - some highest granularity value (not generalized)

generalization - Any value from the VGH domain

Returns:

true if the value generalizes to the specified generalized value

Throws:

java.lang.Exception

getNonLeaves

public java.util.Hashtable<java.lang.String,java.lang.Integer> getNonLeaves()

Get the non-leaf nodes of the VGH

Returns:

A hashtable that maps every non-leaf VGH node to a unique index, assigned by a breadth-first traversal of the VGH

getGeneralizationSequence

public java.lang.String[] getGeneralizationSequence(java.lang.String val)

Get the list of ground-domain values that the provided value represents (applies only to categorical attributes mapped to a discrete domain)

Parameters:

val - some value in the VGH

Returns:

starting with the value itself, returns a vector of generalizations all the way upto the suppression value

getLeafCategories

public int[] getLeafCategories(java.lang.String val)