Functor class for calculating the Tversky Similarity [GSIM] of bitsets. More...

#include <SimilarityFunctors.hpp>

Public Member Functions
	TverskySimilarity (double alpha=0.95, double beta=0.05)
	Constructor. More...

double	operator() (const Util::BitSet &bs1, const Util::BitSet &bs2) const
	Calculates the Tversky Similarity [GSIM] of the bitsets bs1 and bs2. More...

Detailed Description

Functor class for calculating the Tversky Similarity [GSIM] of bitsets.

Constructor & Destructor Documentation

◆ TverskySimilarity()

CDPL::Descr::TverskySimilarity::TverskySimilarity	(	double	alpha = `0.95`,
		double	beta = `0.05`
	)

inline

Constructor.

Parameters

alpha	Weights the contribution of the first bitset.
beta	Weights the contribution of the second bitset.

Member Function Documentation

◆ operator()()

double CDPL::Descr::TverskySimilarity::operator()	(	const Util::BitSet &	bs1,
		const Util::BitSet &	bs2
	)		const

inline

Calculates the Tversky Similarity [GSIM] of the bitsets bs1 and bs2.

The Tversky Similarity $S_{ab}$ is calculated by:

$S_{ab} = \frac{N_{ab}}{\alpha * N_a + \beta * N_b + N_{ab}}$

where $N_{ab}$ is the number of bits that are set in both bitsets, $N_a$ is the number of bits that are only set in the first bitset and $N_b$ is the number of bits that are only set in the second bitset. $\alpha$ and $\beta$ are bitset contribution weighting factors.

The Tversky measure is asymmetric. Setting the parameters $\alpha = \beta = 1.0$ makes it identical to the Tanimoto measure.

If the specified bitsets bs1 and bs2 are of different size, missing bits at the end of the smaller bitset are assumed to be zero.