Similarity.h File Reference

The implementantion of TRACLUS algorithm. More...

#include "Definition.h"

Include dependency graph for Similarity.h:

This graph shows which files directly or indirectly include this file:

Functions
float8 *	EuclideanStart (const Trajectory *trajectoryA, const Trajectory *trajectoryB, float8 *result)
	The function returns the euclidean distance between the first points of the trajectories. More...
PGDLLEXPORT Datum	EuclideanStartV1 (PG_FUNCTION_ARGS)
float8 *	EuclideanEnd (const Trajectory *trajectoryA, const Trajectory *trajectoryB, float8 *result)
	The function returns the euclidean distance between the last points of the trajectories. More...
PGDLLEXPORT Datum	EuclideanEndV1 (PG_FUNCTION_ARGS)
float8 *	EuclideanStartEnd (const Trajectory *trajectoryA, const Trajectory *trajectoryB, float8 *result)
	The function returns the average euclidean distance between the first and last points of the trajectories. More...
PGDLLEXPORT Datum	EuclideanStartEndV1 (PG_FUNCTION_ARGS)
float8 *	Manhattan (const Trajectory *trajectoryA, const Trajectory *trajectoryB, int32 precision, float8 *result)
	The function returns the Manhattan norm between the trajectories. More...
PGDLLEXPORT Datum	ManhattanV1 (PG_FUNCTION_ARGS)
float8 *	Euclidean (const Trajectory *trajectoryA, const Trajectory *trajectoryB, int32 precision, float8 *result)
	The function returns the euclidean norm between the trajectories. More...
PGDLLEXPORT Datum	EuclideanV1 (PG_FUNCTION_ARGS)
float8 *	Chebyshev (const Trajectory *trajectoryA, const Trajectory *trajectoryB, int32 precision, float8 *result)
	The function returns the maximum norm between the trajectories. More...
PGDLLEXPORT Datum	ChebyshevV1 (PG_FUNCTION_ARGS)
float8 *	DISSIM (const Trajectory *trajectoryA, const Trajectory *trajectoryB, int32 precision, float8 *result, float8 *error)
	The function returns DSSIM. More...
PGDLLEXPORT Datum	DISSIMExact (PG_FUNCTION_ARGS)
PGDLLEXPORT Datum	DISSIMApproximate (PG_FUNCTION_ARGS)
float8 *	DTW (const Trajectory *trajectoryA, const Trajectory *trajectoryB, int32 w, float8 w_percentage, int32 lp, float8 *result)
	The function returns the Dynamic Time Warping. More...
PGDLLEXPORT Datum	DTWV1 (PG_FUNCTION_ARGS)
float8 *	LCSS (const Trajectory *trajectoryA, const Trajectory *trajectoryB, int32 w, float8 w_percentage, float8 eps, float8 *result)
	The function returns the Longest Common Subsequence. More...
PGDLLEXPORT Datum	LCSSV1 (PG_FUNCTION_ARGS)
float8 *	EDR (const Trajectory *trajectoryA, const Trajectory *trajectoryB, float8 eps, float8 *result)
	The function returns the EDR. More...
PGDLLEXPORT Datum	EDRV1 (PG_FUNCTION_ARGS)
float8 *	ERP (const Trajectory *trajectoryA, const Trajectory *trajectoryB, int32 lp, float8 *result)
	The function returns the ERP. More...
PGDLLEXPORT Datum	ERPV1 (PG_FUNCTION_ARGS)

Detailed Description

The implementantion of TRACLUS algorithm.

Author

Giannis Kostis (jkost.nosp@m.is@g.nosp@m.mail..nosp@m.com)

Similarity functions between trajectories. For more detailed descritpion look at Distance functions

See also

Distance functions

Trajectory/Similarity.sql

Function Documentation

float8* Chebyshev	(	const Trajectory *	trajectoryA,
		const Trajectory *	trajectoryB,
		int32	precision,
		float8 *	result
	)

The function returns the maximum norm between the trajectories.

Parameters

[in]	trajectoryA	The first trajectory
[in]	trajectoryB	The second trajectory
[in]	precision	The wanted precision
[out]	result	The result

References atInstantTrajectory(), distanceChebyshevPointSP_PointSP(), Trajectory::NrPoints, Trajectory::points, PointST::sp, and PointST::t.

Referenced by ChebyshevV1().

Here is the call graph for this function:

PGDLLEXPORT Datum ChebyshevV1

(

PG_FUNCTION_ARGS

)

References Chebyshev(), and validateInterpolationPrecision().

Here is the call graph for this function:

float8* DISSIM	(	const Trajectory *	trajectoryA,
		const Trajectory *	trajectoryB,
		int32	precision,
		float8 *	result,
		float8 *	error
	)

The function returns DSSIM.

Parameters

[in]	trajectoryA	The first trajectory
[in]	trajectoryB	The second trajectory
[in]	precision	The wanted precision
[out]	result	The result
[out]	error	The error

References atInstantTrajectory(), compute_segment_dissim(), Trajectory::NrPoints, Trajectory::points, and PointST::t.

Referenced by DISSIMApproximate(), and DISSIMExact().

Here is the call graph for this function:

PGDLLEXPORT Datum DISSIMApproximate

(

PG_FUNCTION_ARGS

)

References DISSIM(), and validateInterpolationPrecision().

Here is the call graph for this function:

PGDLLEXPORT Datum DISSIMExact

(

PG_FUNCTION_ARGS

)

References DISSIM(), and validateInterpolationPrecision().

Here is the call graph for this function:

float8* DTW	(	const Trajectory *	trajectoryA,
		const Trajectory *	trajectoryB,
		int32	w,
		float8	w_percentage,
		int32	lp,
		float8 *	result
	)

The function returns the Dynamic Time Warping.

Parameters

[in]	trajectoryA	The first trajectory
[in]	trajectoryB	The second trajectory
[in]	w	the locality constraint
[in]	w_percentage	if w < 0 then someone can use percentage of the the trajectory by defining this value
[in]	lp	the norm to be used (currently only 1 and 2)
[out]	result	The result

References distanceManhattanPointSP_PointSP(), i(), j, min(), Trajectory::NrPoints, Trajectory::points, PointST::sp, and sqDistancePointSP_PointSP().

Referenced by DTWV1().

Here is the call graph for this function:

PGDLLEXPORT Datum DTWV1

(

PG_FUNCTION_ARGS

)

References DTW().

Here is the call graph for this function:

float8* EDR	(	const Trajectory *	trajectoryA,
		const Trajectory *	trajectoryB,
		float8	eps,
		float8 *	result
	)

The function returns the EDR.

Parameters

[in]	trajectoryA	The first trajectory
[in]	trajectoryB	The second trajectory
[in]	eps	the threshold, which determines whether or not two elements match.
[out]	result	The result

References GAP, HIT, i(), j, MISS, Trajectory::NrPoints, Trajectory::points, PointST::sp, PointSP::x, and PointSP::y.

Referenced by EDRV1().

Here is the call graph for this function:

PGDLLEXPORT Datum EDRV1

(

PG_FUNCTION_ARGS

)

References EDR().

Here is the call graph for this function:

float8* ERP	(	const Trajectory *	trajectoryA,
		const Trajectory *	trajectoryB,
		int32	lp,
		float8 *	result
	)

The function returns the ERP.

Parameters

[in]	trajectoryA	The first trajectory
[in]	trajectoryB	The second trajectory
[in]	lp	the norm to be used (currently only 1 and 2)
[out]	result	The result

References distanceManhattanPointSP_PointSP(), distancePointSP_PointSP(), i(), j, Trajectory::NrPoints, Trajectory::points, PointST::sp, PointSP::x, and PointSP::y.

Referenced by ERPV1().

Here is the call graph for this function:

PGDLLEXPORT Datum ERPV1

(

PG_FUNCTION_ARGS

)

References ERP().

Here is the call graph for this function:

float8* Euclidean	(	const Trajectory *	trajectoryA,
		const Trajectory *	trajectoryB,
		int32	precision,
		float8 *	result
	)

The function returns the euclidean norm between the trajectories.

Parameters

[in]	trajectoryA	The first trajectory
[in]	trajectoryB	The second trajectory
[in]	precision	The wanted precision
[out]	result	The result

References atInstantTrajectory(), Trajectory::NrPoints, Trajectory::points, PointST::sp, sqDistancePointSP_PointSP(), and PointST::t.

Referenced by EuclideanV1().

Here is the call graph for this function:

float8* EuclideanEnd	(	const Trajectory *	trajectoryA,
		const Trajectory *	trajectoryB,
		float8 *	result
	)

The function returns the euclidean distance between the last points of the trajectories.

Parameters

[in]	trajectoryA	The first trajectory
[in]	trajectoryB	The second trajectory
[out]	result	The result

References distancePointSP_PointSP(), Trajectory::NrPoints, Trajectory::points, and PointST::sp.

Referenced by EuclideanEndV1().

Here is the call graph for this function:

PGDLLEXPORT Datum EuclideanEndV1

(

PG_FUNCTION_ARGS

)

References EuclideanEnd().

Here is the call graph for this function:

float8* EuclideanStart	(	const Trajectory *	trajectoryA,
		const Trajectory *	trajectoryB,
		float8 *	result
	)

The function returns the euclidean distance between the first points of the trajectories.

Parameters

[in]	trajectoryA	The first trajectory
[in]	trajectoryB	The second trajectory
[out]	result	The result

References distancePointSP_PointSP(), Trajectory::NrPoints, Trajectory::points, and PointST::sp.

Referenced by EuclideanStartV1().

Here is the call graph for this function:

float8* EuclideanStartEnd	(	const Trajectory *	trajectoryA,
		const Trajectory *	trajectoryB,
		float8 *	result
	)

The function returns the average euclidean distance between the first and last points of the trajectories.

Parameters

[in]	trajectoryA	The first trajectory
[in]	trajectoryB	The second trajectory
[out]	result	The result

References distancePointSP_PointSP(), Trajectory::NrPoints, Trajectory::points, and PointST::sp.

Referenced by EuclideanStartEndV1().

Here is the call graph for this function:

PGDLLEXPORT Datum EuclideanStartEndV1

(

PG_FUNCTION_ARGS

)

References EuclideanStartEnd().

Here is the call graph for this function:

PGDLLEXPORT Datum EuclideanStartV1

(

PG_FUNCTION_ARGS

)

References EuclideanStart().

Here is the call graph for this function:

PGDLLEXPORT Datum EuclideanV1

(

PG_FUNCTION_ARGS

)

References Euclidean(), and validateInterpolationPrecision().

Here is the call graph for this function:

float8* LCSS	(	const Trajectory *	trajectoryA,
		const Trajectory *	trajectoryB,
		int32	w,
		float8	w_percentage,
		float8	eps,
		float8 *	result
	)

The function returns the Longest Common Subsequence.

Parameters

[in]	trajectoryA	The first trajectory
[in]	trajectoryB	The second trajectory
[in]	w	the locality constraint
[in]	w_percentage	if w < 0 then someone can use percentage of the the trajectory by defining this value
[in]	eps	the threshold, which determines whether or not two elements match. if eps<=0 or eps>=1 then eps = min(stedev(trajectories))
[out]	result	The result

References i(), j, match(), Trajectory::NrPoints, Trajectory::points, PointST::sp, and stddevTrajectory().

Referenced by LCSSV1().

Here is the call graph for this function:

PGDLLEXPORT Datum LCSSV1

(

PG_FUNCTION_ARGS

)

References LCSS().

Here is the call graph for this function:

float8* Manhattan	(	const Trajectory *	trajectoryA,
		const Trajectory *	trajectoryB,
		int32	precision,
		float8 *	result
	)

The function returns the Manhattan norm between the trajectories.

Parameters

[in]	trajectoryA	The first trajectory
[in]	trajectoryB	The second trajectory
[in]	precision	The wanted precision
[out]	result	The result

References atInstantTrajectory(), distanceManhattanPointSP_PointSP(), Trajectory::NrPoints, Trajectory::points, PointST::sp, and PointST::t.

Referenced by ManhattanV1().

Here is the call graph for this function:

PGDLLEXPORT Datum ManhattanV1

(

PG_FUNCTION_ARGS

)

References Manhattan(), and validateInterpolationPrecision().

Here is the call graph for this function: