Package org.ddogleg.fitting.modelset.distance

Class FitByMedianStatistics<Model,Point>

java.lang.Object

org.ddogleg.fitting.modelset.distance.FitByMedianStatistics<Model,Point>

All Implemented Interfaces:

StatisticalFit<Model,Point>

public class FitByMedianStatistics<Model,Point>
extends Object
implements StatisticalFit<Model,Point>

Computes the median error and prunes points if they have more than the specified percentile error.

Constructor Summary

Constructors

Constructor	Description
FitByMedianStatistics(double pruneThreshold)	Creates a new FitByMedianStatistics.

Method Summary

Modifier and Type	Method	Description
void	computeStatistics()	Computes the statistic error of the model to the data points.
double	getErrorMetric()	Returns the computed statistical error.
void	init(DistanceFromModel<Model,Point> modelError, ArrayDeque<PointIndex<Point>> allPoints)	This is called once to provide access to internal data structures of the owner.
void	prune()	Removes all samples which have an error larger than the specified percentile error.

Methods inherited from class java.lang.Object

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

Constructor Details

FitByMedianStatistics

public FitByMedianStatistics(double pruneThreshold)

Creates a new FitByMedianStatistics.

Parameters:

pruneThreshold - Fraction of samples that are not pruned.

Method Details

init

public void init(DistanceFromModel<Model,Point> modelError,
ArrayDeque<PointIndex<Point>> allPoints)

Description copied from interface: StatisticalFit

This is called once to provide access to internal data structures of the owner.

Specified by:

init in interface StatisticalFit<Model,Point>

Parameters:

modelError - Computes the error between a point and the model

allPoints - Contains all the points which are currently considered part of the model.

computeStatistics

public void computeStatistics()

Description copied from interface: StatisticalFit

Computes the statistic error of the model to the data points.

Specified by:

computeStatistics in interface StatisticalFit<Model,Point>

prune

public void prune()

Removes all samples which have an error larger than the specified percentile error.

Specified by:

prune in interface StatisticalFit<Model,Point>

getErrorMetric

public double getErrorMetric()

Description copied from interface: StatisticalFit

Returns the computed statistical error.

Specified by:

getErrorMetric in interface StatisticalFit<Model,Point>

Returns:

The error.