com.chemaxon.descriptors.metrics

Class TverskyParameters

java.lang.Object

com.chemaxon.descriptors.metrics.TverskyParameters

All Implemented Interfaces:

BinaryVectorComparator, MetricMetadata, Serializable

@Beta
 @PublicAPI
public class TverskyParameters
extends Object
implements MetricMetadata, BinaryVectorComparator, Serializable

Parameters for Tversky index.

Please note that this class is marked with @Beta annotation, so it can be subject of incompatible changes or removal in later releases.

See Also:

http://en.wikipedia.org/wiki/Tversky_index, Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	TverskyParameters.Builder Builder for parameter object.

Nested classes/interfaces inherited from interface com.chemaxon.descriptors.metrics.BinaryVectorComparator

BinaryVectorComparator.Util

Constructor Summary

Constructors

Modifier	Constructor and Description
	TverskyParameters() Instantiate with default values.
	TverskyParameters(double coeffQuery, double coeffTarget) Instantiate with weights.
protected	TverskyParameters(TverskyParameters.Builder builder) Instantiate with builder's actual state.

Method Summary

Modifier and Type	Method and Description
double	calculateBinaryVectorDissimilarity(int bitCount, int targetOnly, int queryOnly, int common) Calculate represented dissimilarity score from binary vector parameters.
double	calculateBinaryVectorSimilarity(int bitCount, int targetOnly, int queryOnly, int common) Calculate represented similarity score from binary vector parameters.
double	getCoeffQuery() Query side coefficient.
double	getCoeffTarget() Target side coefficient.
double	getLowerBoundForBinaryVectors(int bitCount) Possible lowest value of similarity/dissimilarity scores.
UnguardedDissimilarityCalculator<long[]>	getUnguardedDissimilarityCalculator(int bitCount) Get a suitable dissimilarity calculator for unguarded packed long[] representations.
double	getUpperBoundForBinaryVectors(int bitCount) Possible highest value of similarity/dissimilarity scores.
boolean	isDissimilarityZeroIFFEquals() Zero dissimilarity and identity.
boolean	isMetricSpace() Metric space.
boolean	isNonNegative() Non negativeness of the dissimilarity values.
boolean	isSymmetric() Symmetry.
boolean	isTriangeInequalityHolds() Triangle inequality.
static TverskyParameters	of(double coeffQuery, double coeffTarget) Convenience factory function.
String	toString()

Methods inherited from class java.lang.Object

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

Constructor Detail

TverskyParameters

protected TverskyParameters(TverskyParameters.Builder builder)

Instantiate with builder's actual state.

Parameters:

builder - State to represent

TverskyParameters

public TverskyParameters()

Instantiate with default values.

TverskyParameters

public TverskyParameters(double coeffQuery,
                         double coeffTarget)

Instantiate with weights.

Parameters:

coeffQuery - Query side coefficient

coeffTarget - Target side coefficient

Method Detail

isSymmetric

public boolean isSymmetric()

Description copied from interface: MetricMetadata

Symmetry.

Specified by:

isSymmetric in interface MetricMetadata

Returns:

true when d(x,y) = d(y,x)

See Also:

http://en.wikipedia.org/wiki/Metric_space

isNonNegative

public boolean isNonNegative()

Description copied from interface: MetricMetadata

Non negativeness of the dissimilarity values.

Specified by:

isNonNegative in interface MetricMetadata

Returns:

true when d(x,y) >= 0

See Also:

http://en.wikipedia.org/wiki/Metric_space

isDissimilarityZeroIFFEquals

public boolean isDissimilarityZeroIFFEquals()

Description copied from interface: MetricMetadata

Zero dissimilarity and identity.

Specified by:

isDissimilarityZeroIFFEquals in interface MetricMetadata

Returns:

true when dissimilarity is 0 for (and only for) identical descriptors

See Also:

http://en.wikipedia.org/wiki/Metric_space

isTriangeInequalityHolds

public boolean isTriangeInequalityHolds()

Description copied from interface: MetricMetadata

Triangle inequality.

Specified by:

isTriangeInequalityHolds in interface MetricMetadata

Returns:

true when d(x,z) <= d(x,y) + d( y,z)

See Also:

http://en.wikipedia.org/wiki/Metric_space

isMetricSpace

public boolean isMetricSpace()

Description copied from interface: MetricMetadata

Metric space.

Specified by:

isMetricSpace in interface MetricMetadata

Returns:

If this comparator defines a metric space.

See Also:

http://en.wikipedia.org/wiki/Metric_space

calculateBinaryVectorDissimilarity

public double calculateBinaryVectorDissimilarity(int bitCount,
                                                 int targetOnly,
                                                 int queryOnly,
                                                 int common)

Description copied from interface: BinaryVectorComparator

Calculate represented dissimilarity score from binary vector parameters.

Specified by:

calculateBinaryVectorDissimilarity in interface BinaryVectorComparator

Parameters:

bitCount - Length of binary vectors in bits

targetOnly - Number of bit positions where only target contains set bits

queryOnly - Number of bit positions where only query contains set bits

common - Number of bit positions where both query and target contains set bits

Returns:

Dissimilarity score

calculateBinaryVectorSimilarity

public double calculateBinaryVectorSimilarity(int bitCount,
                                              int targetOnly,
                                              int queryOnly,
                                              int common)

Description copied from interface: BinaryVectorComparator

Calculate represented similarity score from binary vector parameters.

Specified by:

calculateBinaryVectorSimilarity in interface BinaryVectorComparator

Parameters:

bitCount - Length of binary vectors in bits

targetOnly - Number of bit positions where only target contains set bits

queryOnly - Number of bit positions where only query contains set bits

common - Number of bit positions where both query and target contains set bits

Returns:

Ssimilarity score

getLowerBoundForBinaryVectors

public double getLowerBoundForBinaryVectors(int bitCount)

Description copied from interface: BinaryVectorComparator

Possible lowest value of similarity/dissimilarity scores.

Specified by:

getLowerBoundForBinaryVectors in interface BinaryVectorComparator

Parameters:

bitCount - Length of binary vectors in bits

Returns:

Possible lowest value of similarity/dissimilarity scores

getUpperBoundForBinaryVectors

public double getUpperBoundForBinaryVectors(int bitCount)

Description copied from interface: BinaryVectorComparator

Possible highest value of similarity/dissimilarity scores.

Specified by:

getUpperBoundForBinaryVectors in interface BinaryVectorComparator

Parameters:

bitCount - Length of binary vectors in bits

Returns:

Possible highest value of similarity/dissimilarity scores

getCoeffQuery

public double getCoeffQuery()

Query side coefficient.

Returns:

Query coefficient

getCoeffTarget

public double getCoeffTarget()

Target side coefficient.

Returns:

Target coefficient

toString

public String toString()

Overrides:

toString in class Object

getUnguardedDissimilarityCalculator

public UnguardedDissimilarityCalculator<long[]> getUnguardedDissimilarityCalculator(int bitCount)

Description copied from interface: BinaryVectorComparator

Get a suitable dissimilarity calculator for unguarded packed long[] representations.

Specified by:

getUnguardedDissimilarityCalculator in interface BinaryVectorComparator

Parameters:

bitCount - Bit count of the fingerprints; always is a multiply of 64

Returns:

Unguarded calculator

of

public static TverskyParameters of(double coeffQuery,
                                   double coeffTarget)

Convenience factory function.

Parameters:

coeffQuery - Query side penalty coefficient

coeffTarget - Target side penalty coefficient

Returns:

representation of given parameters