org.encog.mathutil.error

Class NormalizedError

java.lang.Object

org.encog.mathutil.error.NormalizedError

public class NormalizedError
extends java.lang.Object

A normalized error will generally be in the approximate range between 0 and 1. This allows normalized errors to be compared across multiple datasets. This class implements two types of normalized error calculation: Normalized root-mean-square deviation (NRMSD) and coefficient of variation of the RMSD, CV(RMSD). https://en.wikipedia.org/wiki/Root-mean-square_deviation

Constructor Summary

Constructors

Constructor and Description
NormalizedError(MLDataSet theData) Construct the normalized error calculator.

Method Summary

Modifier and Type	Method and Description
double	calculateNormalizedMean(MLDataSet theData, MLRegression theModel) Calculate the error as the coefficient of variation of the RMSD (CV(RMSD)).
double	calculateNormalizedRange(MLDataSet theData, MLRegression theModel) Calculate the error as the Normalized root-mean-square deviation (NRMSD)

Methods inherited from class java.lang.Object

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

Constructor Detail

NormalizedError

public NormalizedError(MLDataSet theData)

Construct the normalized error calculator.

Parameters:

theData - The dataset to use.

Method Detail

calculateNormalizedMean

public double calculateNormalizedMean(MLDataSet theData,
                                      MLRegression theModel)

Calculate the error as the coefficient of variation of the RMSD (CV(RMSD)).

Parameters:

theData - The dataset to evaluate with.

theModel - The model to evaluate.

Returns:

The CV(RMSD) error.

calculateNormalizedRange

public double calculateNormalizedRange(MLDataSet theData,
                                       MLRegression theModel)

Calculate the error as the Normalized root-mean-square deviation (NRMSD)

Parameters:

theData - The dataset to evaluate with.

theModel - The model to evaluate.

Returns:

The NRMSD error.