|
//-------------------------------------------------------------
// <copyright company=’Microsoft Corporation’>
// Copyright © Microsoft Corporation. All Rights Reserved.
// </copyright>
//-------------------------------------------------------------
// @owner=alexgor, deliant
//=================================================================
// File: StatisticalAnalysis.cs
//
// Namespace: System.Web.UI.WebControls[Windows.Forms].Charting.Formulas
//
// Classes: StatisticalAnalysis
//
// Purpose: This class is used for Statistical Analysis
//
// Reviewed: AG - Apr 1, 2003
//
//===================================================================
using System;
using System.Collections;
#if Microsoft_CONTROL
namespace System.Windows.Forms.DataVisualization.Charting.Formulas
#else
namespace System.Web.UI.DataVisualization.Charting.Formulas
#endif
{
/// <summary>
///
/// </summary>
internal class StatisticalAnalysis : IFormula
{
#region Error strings
// Error strings
//internal string inputArrayStart = "Formula requires";
//internal string inputArrayEnd = "arrays";
#endregion
#region Parameters
/// <summary>
/// Formula Module name
/// </summary>
virtual public string Name { get { return SR.FormulaNameStatisticalAnalysis; } }
#endregion // Parameters
#region Methods
/// <summary>
/// Default constructor
/// </summary>
public StatisticalAnalysis()
{
}
/// <summary>
/// The first method in the module, which converts a formula
/// name to the corresponding private method.
/// </summary>
/// <param name="formulaName">String which represent a formula name</param>
/// <param name="inputValues">Arrays of doubles - Input values</param>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Formula parameters</param>
/// <param name="extraParameterList">Array of strings - Extra Formula parameters from DataManipulator object</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
virtual public void Formula( string formulaName, double [][] inputValues, out double [][] outputValues, string [] parameterList, string [] extraParameterList, out string [][] outLabels )
{
string name;
outLabels = null;
name = formulaName.ToUpper(System.Globalization.CultureInfo.InvariantCulture);
try
{
switch( name )
{
case "TTESTEQUALVARIANCES":
TTest( inputValues, out outputValues, parameterList, out outLabels, true );
break;
case "TTESTUNEQUALVARIANCES":
TTest( inputValues, out outputValues, parameterList, out outLabels, false );
break;
case "TTESTPAIRED":
TTestPaired( inputValues, out outputValues, parameterList, out outLabels );
break;
case "ZTEST":
ZTest( inputValues, out outputValues, parameterList, out outLabels );
break;
case "FTEST":
FTest( inputValues, out outputValues, parameterList, out outLabels );
break;
case "COVARIANCE":
Covariance( inputValues, out outputValues, out outLabels );
break;
case "CORRELATION":
Correlation( inputValues, out outputValues, out outLabels );
break;
case "ANOVA":
Anova( inputValues, out outputValues, parameterList, out outLabels );
break;
case "TDISTRIBUTION":
TDistribution( out outputValues, parameterList, out outLabels );
break;
case "FDISTRIBUTION":
FDistribution( out outputValues, parameterList, out outLabels );
break;
case "NORMALDISTRIBUTION":
NormalDistribution( out outputValues, parameterList, out outLabels );
break;
case "INVERSETDISTRIBUTION":
TDistributionInverse( out outputValues, parameterList, out outLabels );
break;
case "INVERSEFDISTRIBUTION":
FDistributionInverse( out outputValues, parameterList, out outLabels );
break;
case "INVERSENORMALDISTRIBUTION":
NormalDistributionInverse( out outputValues, parameterList, out outLabels );
break;
case "MEAN":
Average( inputValues, out outputValues, out outLabels );
break;
case "VARIANCE":
Variance( inputValues, out outputValues, parameterList, out outLabels );
break;
case "MEDIAN":
Median( inputValues, out outputValues, out outLabels );
break;
case "BETAFUNCTION":
BetaFunction( out outputValues, parameterList, out outLabels );
break;
case "GAMMAFUNCTION":
GammaFunction( out outputValues, parameterList, out outLabels );
break;
default:
outputValues = null;
break;
}
}
catch( IndexOutOfRangeException )
{
throw new InvalidOperationException( SR.ExceptionFormulaInvalidPeriod(name) );
}
catch( OverflowException )
{
throw new InvalidOperationException( SR.ExceptionFormulaNotEnoughDataPoints(name) );
}
}
#endregion // Methods
#region Statistical Tests
/// <summary>
/// Anova test
/// </summary>
/// <param name="inputValues">Arrays of doubles - Input values</param>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void Anova(double [][] inputValues, out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// There is no enough input series
if( inputValues.Length < 3 )
throw new ArgumentException(SR.ExceptionStatisticalAnalysesNotEnoughInputSeries);
outLabels = null;
for( int index = 0; index < inputValues.Length - 1; index++ )
{
if( inputValues[index].Length != inputValues[index+1].Length )
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidAnovaTest);
}
// Alpha value
double alpha;
try
{
alpha = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidAlphaValue);
}
if( alpha < 0 || alpha > 1 )
{
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesInvalidAlphaValue);
}
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [10];
// X
outputValues[0] = new double [10];
// Y
outputValues[1] = new double [10];
int m = inputValues.Length - 1;
int n = inputValues[0].Length;
double [] average = new double[ m ];
double [] variance = new double[ m ];
// Find averages
for( int group = 0; group < m; group++ )
{
average[group] = Mean( inputValues[group+1] );
}
// Find variances
for( int group = 0; group < m; group++ )
{
variance[group] = Variance( inputValues[group+1], true );
}
// Total Average ( for all groups )
double averageTotal = Mean( average );
// Total Sample Variance
double totalS = 0;
foreach( double avr in average )
{
totalS += ( avr - averageTotal ) * ( avr - averageTotal );
}
totalS /= ( m - 1 );
// Group Sample Variance
double groupS = Mean( variance );
// F Statistica
double f = totalS * ( n ) / groupS;
// ****************************************
// Sum of Squares
// ****************************************
// Grend Total Average
double grandTotalAverage = 0;
for( int group = 0; group < m; group++ )
{
foreach( double point in inputValues[group+1] )
{
grandTotalAverage += point;
}
}
grandTotalAverage /= ( m * n );
// Treatment Sum of Squares
double trss = 0;
for( int group = 0; group < m; group++ )
{
trss += ( average[group] - grandTotalAverage ) * ( average[group] - grandTotalAverage );
}
trss *= n;
// Error Sum of Squares
double erss = 0;
for( int group = 0; group < m; group++ )
{
foreach( double point in inputValues[group+1] )
{
erss += ( point - average[group] ) * ( point - average[group] );
}
}
outLabels[0][0] = SR.LabelStatisticalSumOfSquaresBetweenGroups;
outputValues[0][0] = 1;
outputValues[1][0] = trss;
outLabels[0][1] = SR.LabelStatisticalSumOfSquaresWithinGroups;
outputValues[0][1] = 2;
outputValues[1][1] = erss;
outLabels[0][2] = SR.LabelStatisticalSumOfSquaresTotal;
outputValues[0][2] = 3;
outputValues[1][2] = trss + erss;
outLabels[0][3] = SR.LabelStatisticalDegreesOfFreedomBetweenGroups;
outputValues[0][3] = 4;
outputValues[1][3] = m - 1;
outLabels[0][4] = SR.LabelStatisticalDegreesOfFreedomWithinGroups;
outputValues[0][4] = 5;
outputValues[1][4] = m * ( n - 1 );
outLabels[0][5] = SR.LabelStatisticalDegreesOfFreedomTotal;
outputValues[0][5] = 6;
outputValues[1][5] = m * n - 1;
outLabels[0][6] = SR.LabelStatisticalMeanSquareVarianceBetweenGroups;
outputValues[0][6] = 7;
outputValues[1][6] = trss / ( m - 1 );
outLabels[0][7] = SR.LabelStatisticalMeanSquareVarianceWithinGroups;
outputValues[0][7] = 8;
outputValues[1][7] = erss / ( m * ( n - 1 ) );
outLabels[0][8] = SR.LabelStatisticalFRatio;
outputValues[0][8] = 9;
outputValues[1][8] = f;
outLabels[0][9] = SR.LabelStatisticalFCriteria;
outputValues[0][9] = 10;
outputValues[1][9] = FDistributionInverse( alpha, m - 1, m * ( n - 1 ) );
}
/// <summary>
/// Correlation measure the relationship between two data sets that
/// are scaled to be independent of the unit of measurement. The
/// population correlation calculation returns the covariance
/// of two data sets divided by the product of their standard
/// deviations: You can use the Correlation to determine whether two
/// ranges of data move together — that is, whether large values of
/// one set are associated with large values of the other
/// (positive correlation), whether small values of one set are
/// associated with large values of the other (negative correlation),
/// or whether values in both sets are unrelated (correlation
/// near zero).
/// </summary>
/// <param name="inputValues">Arrays of doubles - Input values</param>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void Correlation(double [][] inputValues, out double [][] outputValues, out string [][] outLabels )
{
// There is no enough input series
if( inputValues.Length != 3 )
throw new ArgumentException( SR.ExceptionPriceIndicatorsFormulaRequiresTwoArrays);
outLabels = null;
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
// Find Covariance.
double covar = Covar( inputValues[1], inputValues[2] );
double varianceX = Variance( inputValues[1], false );
double varianceY = Variance( inputValues[2], false );
// Correlation
double correl = covar / Math.Sqrt( varianceX * varianceY );
outLabels[0][0] = SR.LabelStatisticalCorrelation;
outputValues[0][0] = 1;
outputValues[1][0] = correl;
}
/// <summary>
/// Returns covariance, the average of the products of deviations
/// for each data point pair. Use covariance to determine the
/// relationship between two data sets. For example, you can
/// examine whether greater income accompanies greater
/// levels of education.
/// </summary>
/// <param name="inputValues">Arrays of doubles - Input values</param>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void Covariance(double [][] inputValues, out double [][] outputValues, out string [][] outLabels )
{
// There is no enough input series
if( inputValues.Length != 3 )
throw new ArgumentException( SR.ExceptionPriceIndicatorsFormulaRequiresTwoArrays);
outLabels = null;
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
// Find Covariance.
double covar = Covar( inputValues[1], inputValues[2] );
outLabels[0][0] = SR.LabelStatisticalCovariance;
outputValues[0][0] = 1;
outputValues[1][0] = covar;
}
/// <summary>
/// Returns the result of an F-test. An F-test returns the one-tailed
/// probability that the variances in array1 and array2 are not
/// significantly different. Use this function to determine
/// whether two samples have different variances. For example,
/// given test scores from public and private schools, you can
/// test whether these schools have different levels of diversity.
/// </summary>
/// <param name="inputValues">Arrays of doubles - Input values</param>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void FTest(double [][] inputValues, out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// There is no enough input series
if( inputValues.Length != 3 )
throw new ArgumentException( SR.ExceptionPriceIndicatorsFormulaRequiresTwoArrays);
outLabels = null;
double alpha;
// The number of data points has to be > 1.
CheckNumOfPoints( inputValues );
// Alpha value
try
{
alpha = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidAlphaValue);
}
if( alpha < 0 || alpha > 1 )
{
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesInvalidAlphaValue);
}
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [7];
// X
outputValues[0] = new double [7];
// Y
outputValues[1] = new double [7];
// Find Variance of the first group
double variance1 = Variance( inputValues[1], true );
// Find Variance of the second group
double variance2 = Variance( inputValues[2], true );
// Find Mean of the first group
double mean1 = Mean( inputValues[1] );
// Find Mean of the second group
double mean2 = Mean( inputValues[2] );
// F Value
double valueF = variance1 / variance2;
if( variance2 == 0 )
{
throw new InvalidOperationException(SR.ExceptionStatisticalAnalysesZeroVariance);
}
// The way to find a left critical value is to reversed the degrees of freedom,
// look up the right critical value, and then take the reciprocal of this value.
// For example, the critical value with 0.05 on the left with 12 numerator and 15
// denominator degrees of freedom is found of taking the reciprocal of the critical
// value with 0.05 on the right with 15 numerator and 12 denominator degrees of freedom.
// Avoiding Left Critical Values. Since the left critical values are a pain to calculate,
// they are often avoided altogether. This is the procedure followed in the textbook.
// You can force the F test into a right tail test by placing the sample with the large
// variance in the numerator and the smaller variance in the denominator. It does not
// matter which sample has the larger sample size, only which sample has the larger
// variance. The numerator degrees of freedom will be the degrees of freedom for
// whichever sample has the larger variance (since it is in the numerator) and the
// denominator degrees of freedom will be the degrees of freedom for whichever sample
// has the smaller variance (since it is in the denominator).
bool lessOneF = valueF <= 1;
double fDistInv;
double fDist;
if( lessOneF )
{
fDistInv = FDistributionInverse( 1 - alpha, inputValues[1].Length - 1, inputValues[2].Length - 1 );
fDist = 1 - FDistribution( valueF, inputValues[1].Length - 1, inputValues[2].Length - 1 );
}
else
{
fDistInv = FDistributionInverse( alpha, inputValues[1].Length - 1, inputValues[2].Length - 1 );
fDist = FDistribution( valueF, inputValues[1].Length - 1, inputValues[2].Length - 1 );
}
outLabels[0][0] = SR.LabelStatisticalTheFirstGroupMean;
outputValues[0][0] = 1;
outputValues[1][0] = mean1;
outLabels[0][1] = SR.LabelStatisticalTheSecondGroupMean;
outputValues[0][1] = 2;
outputValues[1][1] = mean2;
outLabels[0][2] = SR.LabelStatisticalTheFirstGroupVariance;
outputValues[0][2] = 3;
outputValues[1][2] = variance1;
outLabels[0][3] = SR.LabelStatisticalTheSecondGroupVariance;
outputValues[0][3] = 4;
outputValues[1][3] = variance2;
outLabels[0][4] = SR.LabelStatisticalFValue;
outputValues[0][4] = 5;
outputValues[1][4] = valueF;
outLabels[0][5] = SR.LabelStatisticalPFLessEqualSmallFOneTail;
outputValues[0][5] = 6;
outputValues[1][5] = fDist;
outLabels[0][6] = SR.LabelStatisticalFCriticalValueOneTail;
outputValues[0][6] = 7;
outputValues[1][6] = fDistInv;
}
/// <summary>
/// Returns the two-tailed P-value of a z-test. The z-test
/// generates a standard score for x with respect to the data set,
/// array, and returns the two-tailed probability for the
/// normal distribution. You can use this function to assess
/// the likelihood that a particular observation is drawn
/// from a particular population.
/// </summary>
/// <param name="inputValues">Arrays of doubles - Input values</param>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void ZTest(double [][] inputValues, out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// There is no enough input series
if( inputValues.Length != 3 )
throw new ArgumentException( SR.ExceptionPriceIndicatorsFormulaRequiresTwoArrays);
// The number of data points has to be > 1.
CheckNumOfPoints( inputValues );
outLabels = null;
double variance1;
double variance2;
double alpha;
double HypothesizedMeanDifference;
// Find Hypothesized Mean Difference parameter
try
{
HypothesizedMeanDifference = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidMeanDifference);
}
if( HypothesizedMeanDifference < 0.0 )
{
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesNegativeMeanDifference);
}
// Find variance of the first group
try
{
variance1 = double.Parse( parameterList[1], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidVariance);
}
// Find variance of the second group
try
{
variance2 = double.Parse( parameterList[2], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidVariance);
}
// Alpha value
try
{
alpha = double.Parse( parameterList[3], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidAlphaValue);
}
if( alpha < 0 || alpha > 1 )
{
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesInvalidAlphaValue);
}
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [9];
// X
outputValues[0] = new double [9];
// Y
outputValues[1] = new double [9];
// Find Mean of the first group
double mean1 = Mean( inputValues[1] );
// Find Mean of the second group
double mean2 = Mean( inputValues[2] );
double dev = Math.Sqrt( variance1 / inputValues[1].Length + variance2 / inputValues[2].Length );
// Z Value
double valueZ = ( mean1 - mean2 - HypothesizedMeanDifference ) / dev;
double normalDistTwoInv = NormalDistributionInverse( 1 - alpha / 2 );
double normalDistOneInv = NormalDistributionInverse( 1 - alpha);
double normalDistOne;
double normalDistTwo;
if( valueZ < 0.0 )
{
normalDistOne = NormalDistribution( valueZ );
}
else
{
normalDistOne = 1.0 - NormalDistribution( valueZ );
}
normalDistTwo = 2.0 * normalDistOne;
outLabels[0][0] = SR.LabelStatisticalTheFirstGroupMean;
outputValues[0][0] = 1;
outputValues[1][0] = mean1;
outLabels[0][1] = SR.LabelStatisticalTheSecondGroupMean;
outputValues[0][1] = 2;
outputValues[1][1] = mean2;
outLabels[0][2] = SR.LabelStatisticalTheFirstGroupVariance;
outputValues[0][2] = 3;
outputValues[1][2] = variance1;
outLabels[0][3] = SR.LabelStatisticalTheSecondGroupVariance;
outputValues[0][3] = 4;
outputValues[1][3] = variance2;
outLabels[0][4] = SR.LabelStatisticalZValue;
outputValues[0][4] = 5;
outputValues[1][4] = valueZ;
outLabels[0][5] = SR.LabelStatisticalPZLessEqualSmallZOneTail;
outputValues[0][5] = 6;
outputValues[1][5] = normalDistOne;
outLabels[0][6] = SR.LabelStatisticalZCriticalValueOneTail;
outputValues[0][6] = 7;
outputValues[1][6] = normalDistOneInv;
outLabels[0][7] = SR.LabelStatisticalPZLessEqualSmallZTwoTail;
outputValues[0][7] = 8;
outputValues[1][7] = normalDistTwo;
outLabels[0][8] = SR.LabelStatisticalZCriticalValueTwoTail;
outputValues[0][8] = 9;
outputValues[1][8] = normalDistTwoInv;
}
/// <summary>
/// Returns the two-tailed P-value of a z-test. The z-test
/// generates a standard score for x with respect to the data set,
/// array, and returns the two-tailed probability for the
/// normal distribution. You can use this function to assess
/// the likelihood that a particular observation is drawn
/// from a particular population.
/// </summary>
/// <param name="inputValues">Arrays of doubles - Input values</param>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
/// <param name="equalVariances">True if Variances are equal.</param>
private void TTest(double [][] inputValues, out double [][] outputValues, string [] parameterList, out string [][] outLabels, bool equalVariances )
{
// There is no enough input series
if( inputValues.Length != 3 )
throw new ArgumentException( SR.ExceptionPriceIndicatorsFormulaRequiresTwoArrays);
outLabels = null;
double variance1;
double variance2;
double alpha;
double HypothesizedMeanDifference;
// Find Hypothesized Mean Difference parameter
try
{
HypothesizedMeanDifference = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidMeanDifference);
}
if( HypothesizedMeanDifference < 0.0 )
{
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesNegativeMeanDifference);
}
// Alpha value
try
{
alpha = double.Parse( parameterList[1], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidAlphaValue);
}
if( alpha < 0 || alpha > 1 )
{
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesInvalidAlphaValue);
}
// The number of data points has to be > 1.
CheckNumOfPoints( inputValues );
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [10];
// X
outputValues[0] = new double [10];
// Y
outputValues[1] = new double [10];
// Find Mean of the first group
double mean1 = Mean( inputValues[1] );
// Find Mean of the second group
double mean2 = Mean( inputValues[2] );
variance1 = Variance( inputValues[1], true );
variance2 = Variance( inputValues[2], true );
double s;
double T;
int freedom;
if( equalVariances )
{
freedom = inputValues[1].Length + inputValues[2].Length - 2;
// S value
s = ( ( inputValues[1].Length - 1 ) * variance1 + ( inputValues[2].Length - 1 ) * variance2 ) / ( inputValues[1].Length + inputValues[2].Length - 2 );
// T value
T = ( mean1 - mean2 - HypothesizedMeanDifference ) / ( Math.Sqrt( s * ( 1.0 / inputValues[1].Length + 1.0 / inputValues[2].Length ) ) );
}
else
{
double m = inputValues[1].Length;
double n = inputValues[2].Length;
double s1 = variance1;
double s2 = variance2;
double f = ( s1 / m + s2 / n ) * ( s1 / m + s2 / n ) / ( ( s1 / m ) * ( s1 / m ) / ( m - 1 ) + ( s2 / n ) * ( s2 / n ) / ( n - 1 ) );
freedom = (int)Math.Round(f);
s = Math.Sqrt( variance1 / inputValues[1].Length + variance2 / inputValues[2].Length );
// Z Value
T = ( mean1 - mean2 - HypothesizedMeanDifference ) / s;
}
double TDistTwoInv = StudentsDistributionInverse( alpha , freedom );
bool more50 = alpha > 0.5;
if( more50 )
{
alpha = 1 - alpha;
}
double TDistOneInv = StudentsDistributionInverse( alpha * 2.0, freedom );
if( more50 )
{
TDistOneInv *= -1.0;
}
double TDistTwo = StudentsDistribution( T, freedom, false );
double TDistOne = StudentsDistribution( T, freedom, true );
outLabels[0][0] = SR.LabelStatisticalTheFirstGroupMean;
outputValues[0][0] = 1;
outputValues[1][0] = mean1;
outLabels[0][1] = SR.LabelStatisticalTheSecondGroupMean;
outputValues[0][1] = 2;
outputValues[1][1] = mean2;
outLabels[0][2] = SR.LabelStatisticalTheFirstGroupVariance;
outputValues[0][2] = 3;
outputValues[1][2] = variance1;
outLabels[0][3] = SR.LabelStatisticalTheSecondGroupVariance;
outputValues[0][3] = 4;
outputValues[1][3] = variance2;
outLabels[0][4] = SR.LabelStatisticalTValue;
outputValues[0][4] = 5;
outputValues[1][4] = T;
outLabels[0][5] = SR.LabelStatisticalDegreeOfFreedom;
outputValues[0][5] = 6;
outputValues[1][5] = freedom;
outLabels[0][6] = SR.LabelStatisticalPTLessEqualSmallTOneTail;
outputValues[0][6] = 7;
outputValues[1][6] = TDistOne;
outLabels[0][7] = SR.LabelStatisticalSmallTCrititcalOneTail;
outputValues[0][7] = 8;
outputValues[1][7] = TDistOneInv;
outLabels[0][8] = SR.LabelStatisticalPTLessEqualSmallTTwoTail;
outputValues[0][8] = 9;
outputValues[1][8] = TDistTwo;
outLabels[0][9] = SR.LabelStatisticalSmallTCrititcalTwoTail;
outputValues[0][9] = 10;
outputValues[1][9] = TDistTwoInv;
}
/// <summary>
/// Returns the two-tailed P-value of a z-test. The z-test
/// generates a standard score for x with respect to the data set,
/// array, and returns the two-tailed probability for the
/// normal distribution. You can use this function to assess
/// the likelihood that a particular observation is drawn
/// from a particular population.
/// </summary>
/// <param name="inputValues">Arrays of doubles - Input values</param>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void TTestPaired(double [][] inputValues, out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// There is no enough input series
if( inputValues.Length != 3 )
throw new ArgumentException( SR.ExceptionPriceIndicatorsFormulaRequiresTwoArrays);
if( inputValues[1].Length != inputValues[2].Length )
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidVariableRanges);
outLabels = null;
double variance;
double alpha;
double HypothesizedMeanDifference;
int freedom;
// Find Hypothesized Mean Difference parameter
try
{
HypothesizedMeanDifference = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidMeanDifference);
}
if( HypothesizedMeanDifference < 0.0 )
{
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesNegativeMeanDifference);
}
// Alpha value
try
{
alpha = double.Parse( parameterList[1], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidAlphaValue);
}
if( alpha < 0 || alpha > 1 )
{
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesInvalidAlphaValue);
}
// The number of data points has to be > 1.
CheckNumOfPoints( inputValues );
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [10];
// X
outputValues[0] = new double [10];
// Y
outputValues[1] = new double [10];
double [] difference = new double[inputValues[1].Length];
for( int item = 0; item < inputValues[1].Length; item++ )
{
difference[item] = inputValues[1][item] - inputValues[2][item];
}
// Find Mean of the second group
double mean = Mean( difference );
variance = Math.Sqrt( Variance( difference, true ) );
double T = ( Math.Sqrt( inputValues[1].Length ) * ( mean - HypothesizedMeanDifference ) ) / variance;
freedom = inputValues[1].Length - 1;
double TDistTwoInv = StudentsDistributionInverse( alpha , freedom );
double TDistOneInv = alpha <= 0.5 ? StudentsDistributionInverse(2 * alpha, freedom) : double.NaN;
double TDistTwo = StudentsDistribution( T, freedom, false );
double TDistOne = StudentsDistribution( T, freedom, true );
outLabels[0][0] = SR.LabelStatisticalTheFirstGroupMean;
outputValues[0][0] = 1;
outputValues[1][0] = Mean(inputValues[1]);
outLabels[0][1] = SR.LabelStatisticalTheSecondGroupMean;
outputValues[0][1] = 2;
outputValues[1][1] = Mean(inputValues[2]);
outLabels[0][2] = SR.LabelStatisticalTheFirstGroupVariance;
outputValues[0][2] = 3;
outputValues[1][2] = Variance(inputValues[1],true);
outLabels[0][3] = SR.LabelStatisticalTheSecondGroupVariance;
outputValues[0][3] = 4;
outputValues[1][3] = Variance(inputValues[2],true);
outLabels[0][4] = SR.LabelStatisticalTValue;
outputValues[0][4] = 5;
outputValues[1][4] = T;
outLabels[0][5] = SR.LabelStatisticalDegreeOfFreedom;
outputValues[0][5] = 6;
outputValues[1][5] = freedom;
outLabels[0][6] = SR.LabelStatisticalPTLessEqualSmallTOneTail;
outputValues[0][6] = 7;
outputValues[1][6] = TDistOne;
outLabels[0][7] = SR.LabelStatisticalSmallTCrititcalOneTail;
outputValues[0][7] = 8;
outputValues[1][7] = TDistOneInv;
outLabels[0][8] = SR.LabelStatisticalPTLessEqualSmallTTwoTail;
outputValues[0][8] = 9;
outputValues[1][8] = TDistTwo;
outLabels[0][9] = SR.LabelStatisticalSmallTCrititcalTwoTail;
outputValues[0][9] = 10;
outputValues[1][9] = TDistTwoInv;
}
#endregion // Statistical Tests
#region Public distributions
/// <summary>
/// Returns the Percentage Points (probability) for the Student
/// t-distribution. The t-distribution is used in the hypothesis
/// testing of small sample data sets. Use this function in place
/// of a table of critical values for the t-distribution.
/// </summary>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void TDistribution(out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// T value value
double tValue;
try
{
tValue = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidTValue);
}
// DegreeOfFreedom
int freedom;
try
{
freedom = int.Parse( parameterList[1], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidDegreeOfFreedom);
}
// One Tailed distribution
int oneTailed;
try
{
oneTailed = int.Parse( parameterList[2], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidTailedParameter);
}
outLabels = null;
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
outLabels[0][0] = SR.LabelStatisticalProbability;
outputValues[0][0] = 1;
outputValues[1][0] = StudentsDistribution( tValue, freedom, oneTailed == 1 );
}
/// <summary>
/// Returns the F probability distribution. You can use
/// this function to determine whether two data sets have
/// different degrees of diversity. For example, you can
/// examine test scores given to men and women entering
/// high school and determine if the variability in the
/// females is different from that found in the males.
/// </summary>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void FDistribution(out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// F value value
double fValue;
try
{
fValue = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidTValue);
}
// Degree Of Freedom 1
int freedom1;
try
{
freedom1 = int.Parse( parameterList[1], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidDegreeOfFreedom);
}
// Degree Of Freedom 2
int freedom2;
try
{
freedom2 = int.Parse( parameterList[2], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidDegreeOfFreedom);
}
outLabels = null;
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
outLabels[0][0] = SR.LabelStatisticalProbability;
outputValues[0][0] = 1;
outputValues[1][0] = FDistribution( fValue, freedom1, freedom2 );
}
/// <summary></summary>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void NormalDistribution(out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// F value value
double zValue;
try
{
zValue = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidZValue);
}
outLabels = null;
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
outLabels[0][0] = SR.LabelStatisticalProbability;
outputValues[0][0] = 1;
outputValues[1][0] = this.NormalDistribution( zValue );
}
/// <summary>
/// Returns the t-value of the Student's t-distribution
/// as a function of the probability and the degrees
/// of freedom.
/// </summary>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void TDistributionInverse(out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// T value value
double probability;
try
{
probability = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidProbabilityValue);
}
// DegreeOfFreedom
int freedom;
try
{
freedom = int.Parse( parameterList[1], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidDegreeOfFreedom);
}
outLabels = null;
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
outLabels[0][0] = SR.LabelStatisticalProbability;
outputValues[0][0] = 1;
outputValues[1][0] = StudentsDistributionInverse( probability, freedom );
}
/// <summary>
/// Returns the inverse of the F probability distribution.
/// If p = FDIST(x,...), then FINV(p,...) = x. The F distribution
/// can be used in an F-test that compares the degree of
/// variability in two data sets. For example, you can analyze
/// income distributions in the United States and Canada to
/// determine whether the two ---- have a similar degree
/// of diversity.
/// </summary>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void FDistributionInverse(out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// Probability value value
double probability;
try
{
probability = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidProbabilityValue);
}
// Degree Of Freedom 1
int freedom1;
try
{
freedom1 = int.Parse( parameterList[1], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidDegreeOfFreedom);
}
// Degree Of Freedom 2
int freedom2;
try
{
freedom2 = int.Parse( parameterList[2], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidDegreeOfFreedom);
}
outLabels = null;
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
outLabels[0][0] = SR.LabelStatisticalProbability;
outputValues[0][0] = 1;
outputValues[1][0] = FDistributionInverse( probability, freedom1, freedom2 );
}
/// <summary>
/// Returns the inverse of the standard normal
/// cumulative distribution. The distribution
/// has a mean of zero and a standard deviation
/// of one.
/// </summary>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void NormalDistributionInverse(out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// Alpha value value
double alpha;
try
{
alpha = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidAlphaValue);
}
outLabels = null;
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
outLabels[0][0] = SR.LabelStatisticalProbability;
outputValues[0][0] = 1;
outputValues[1][0] = this.NormalDistributionInverse( alpha );
}
#endregion
#region Utility Statistical Functions
/// <summary>
/// Check number of data points. The number should be greater then 1.
/// </summary>
/// <param name="inputValues">Input series</param>
private void CheckNumOfPoints( double [][] inputValues )
{
if( inputValues[1].Length < 2 )
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesNotEnoughDataPoints);
}
if( inputValues.Length > 2 )
{
if( inputValues[2].Length < 2 )
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesNotEnoughDataPoints);
}
}
}
/// <summary>
/// Returns covariance, the average of the products of deviations
/// for each data point pair. Use covariance to determine the
/// relationship between two data sets. For example, you can
/// examine whether greater income accompanies greater
/// levels of education.
/// </summary>
/// <param name="arrayX">First data set from X random variable.</param>
/// <param name="arrayY">Second data set from Y random variable.</param>
/// <returns>Returns covariance</returns>
private double Covar( double [] arrayX, double [] arrayY )
{
// Check the number of data points
if( arrayX.Length != arrayY.Length )
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesCovariance);
}
double [] arrayXY = new double[arrayX.Length];
// Find XY
for( int index = 0; index < arrayX.Length; index++ )
{
arrayXY[index] = arrayX[index] * arrayY[index];
}
// Find means
double meanXY = Mean( arrayXY );
double meanX = Mean( arrayX );
double meanY = Mean( arrayY );
// return covariance
return meanXY - meanX * meanY;
}
/// <summary>
/// Returns the natural logarithm of the gamma function, G(x).
/// </summary>
/// <param name="n">The value for which you want to calculate gamma function.</param>
/// <returns>Returns the natural logarithm of the gamma function.</returns>
private double GammLn( double n )
{
double x;
double y;
double tmp;
double sum;
double [] cof = {76.18009172947146, -86.50532032941677, 24.01409824083091, -1.231739572450155, 0.1208650973866179e-2, -0.5395239384953e-5};
if( n < 0 )
{
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesGammaBetaNegativeParameters);
}
// Iterative method for Gamma function
y = x = n;
tmp = x + 5.5;
tmp -= ( x + 0.5 ) * Math.Log( tmp );
sum = 1.000000000190015;
for( int item = 0; item <=5; item++ )
{
sum += cof[item] / ++y;
}
return -tmp + Math.Log( 2.5066282746310005 * sum / x );
}
/// <summary>
/// Calculates Beta function
/// </summary>
/// <param name="m">First parameter for beta function</param>
/// <param name="n">Second parameter for beta function</param>
/// <returns>returns beta function</returns>
private double BetaFunction( double m, double n )
{
return Math.Exp( GammLn( m ) + GammLn( n ) - GammLn( m + n ) );
}
/// <summary>
/// Used by betai: Evaluates continued fraction for
/// incomplete beta function by modified Lentz’s
/// </summary>
/// <param name="a">Beta incomplete parameter</param>
/// <param name="b">Beta incomplete parameter</param>
/// <param name="x">Beta incomplete parameter</param>
/// <returns>Value used for Beta incomplete function</returns>
private double BetaCF( double a, double b, double x )
{
int MAXIT = 100;
double EPS = 3.0e-7;
double FPMIN = 1.0e-30;
int m,m2;
double aa,c,d,del,h,qab,qam,qap;
qab = a + b;
qap= a + 1.0;
qam = a - 1.0;
c = 1.0;
d = 1.0 - qab * x / qap;
if ( Math.Abs(d) < FPMIN ) d=FPMIN;
d = 1.0 / d;
h = d;
// Numerical approximation for Beta incomplete function
for( m=1; m<=MAXIT; m++ )
{
m2 = 2*m;
aa = m*(b-m)*x/((qam+m2)*(a+m2));
// Find d coeficient
d = 1.0 + aa*d;
if( Math.Abs(d) < FPMIN ) d=FPMIN;
// Find c coeficient
c = 1.0 + aa / c;
if( Math.Abs(c) < FPMIN ) c = FPMIN;
// Find d coeficient
d = 1.0 / d;
// Find h coeficient
h *= d*c;
aa = -(a+m)*(qab+m)*x/((a+m2)*(qap+m2));
// Recalc d coeficient
d=1.0+aa*d;
if (Math.Abs(d) < FPMIN) d=FPMIN;
// Recalc c coeficient
c=1.0+aa/c;
if (Math.Abs(c) < FPMIN) c=FPMIN;
// Recalc d coeficient
d=1.0/d;
del=d*c;
// Recalc h coeficient
h *= del;
if (Math.Abs(del-1.0) < EPS)
{
break;
}
}
if (m > MAXIT)
{
throw new InvalidOperationException(SR.ExceptionStatisticalAnalysesIncompleteBetaFunction);
}
return h;
}
/// <summary>
/// Standard normal density function
/// </summary>
/// <param name="t">T Value</param>
/// <returns>Standard normal density</returns>
private double NormalDistributionFunction(double t)
{
return 0.398942280401433 * Math.Exp( -t * t / 2 );
}
/// <summary>
/// Returns the incomplete beta function Ix(a, b).
/// </summary>
/// <param name="a">Beta incomplete parameter</param>
/// <param name="b">Beta incomplete parameter</param>
/// <param name="x">Beta incomplete parameter</param>
/// <returns>Beta Incomplete value</returns>
private double BetaIncomplete( double a, double b, double x )
{
double bt;
if (x < 0.0 || x > 1.0)
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesInvalidInputParameter);
if (x == 0.0 || x == 1.0)
{
bt = 0.0;
}
else
{ // Factors in front of the continued fraction.
bt = Math.Exp(GammLn(a + b) - GammLn(a) - GammLn(b) + a * Math.Log(x) + b * Math.Log(1.0 - x));
}
if (x < (a + 1.0) / (a + b + 2.0))
{ //Use continued fraction directly.
return bt * BetaCF(a, b, x) / a;
}
else
{ // Use continued fraction after making the symmetry transformation.
return 1.0 - bt * BetaCF(b, a, 1.0 - x) / b;
}
}
#endregion // Utility Statistical Functions
#region Statistical Parameters
/// <summary>
/// Returns the average (arithmetic mean) of the arguments.
/// </summary>
/// <param name="inputValues">Arrays of doubles - Input values</param>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void Average(double [][] inputValues, out double [][] outputValues, out string [][] outLabels )
{
outLabels = null;
// Invalid number of data series
if( inputValues.Length != 2 )
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidSeriesNumber);
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
outLabels[0][0] = SR.LabelStatisticalAverage;
outputValues[0][0] = 1;
outputValues[1][0] = Mean( inputValues[1] );
}
/// <summary>
/// Calculates variance
/// </summary>
/// <param name="inputValues">Arrays of doubles - Input values</param>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void Variance(double [][] inputValues, out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// Sample Variance value
bool sampleVariance;
try
{
sampleVariance = bool.Parse( parameterList[0] );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidVariance);
}
CheckNumOfPoints(inputValues);
// Invalid number of data series
if( inputValues.Length != 2 )
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidSeriesNumber);
outLabels = null;
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
outLabels[0][0] = SR.LabelStatisticalVariance;
outputValues[0][0] = 1;
outputValues[1][0] = Variance( inputValues[1], sampleVariance );
}
/// <summary>
/// Calculates Median
/// </summary>
/// <param name="inputValues">Arrays of doubles - Input values</param>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void Median(double [][] inputValues, out double [][] outputValues, out string [][] outLabels )
{
outLabels = null;
// Invalid number of data series
if( inputValues.Length != 2 )
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidSeriesNumber);
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
outLabels[0][0] = SR.LabelStatisticalMedian;
outputValues[0][0] = 1;
outputValues[1][0] = Median( inputValues[1] );
}
/// <summary>
/// Calculates Beta Function
/// </summary>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void BetaFunction(out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// Degree of freedom
double m;
try
{
m = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidDegreeOfFreedom);
}
// Degree of freedom
double n;
try
{
n = double.Parse( parameterList[1], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidDegreeOfFreedom);
}
outLabels = null;
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
outLabels[0][0] = SR.LabelStatisticalBetaFunction;
outputValues[0][0] = 1;
outputValues[1][0] = BetaFunction( m, n );
}
/// <summary>
/// Calculates Gamma Function
/// </summary>
/// <param name="outputValues">Arrays of doubles - Output values</param>
/// <param name="parameterList">Array of strings - Parameters</param>
/// <param name="outLabels">Array of strings - Used for Labels. Description for output results.</param>
private void GammaFunction(out double [][] outputValues, string [] parameterList, out string [][] outLabels )
{
// Degree of freedom
double m;
try
{
m = double.Parse( parameterList[0], System.Globalization.CultureInfo.InvariantCulture );
}
catch(System.Exception)
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidInputParameter);
}
if( m < 0 )
{
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesGammaBetaNegativeParameters);
}
outLabels = null;
// Output arrays
outputValues = new double [2][];
// Output Labels
outLabels = new string [1][];
// Parameters description
outLabels[0] = new string [1];
// X
outputValues[0] = new double [1];
// Y
outputValues[1] = new double [1];
outLabels[0][0] = SR.LabelStatisticalGammaFunction;
outputValues[0][0] = 1;
outputValues[1][0] = Math.Exp( GammLn( m ) );
}
/// <summary>
/// Sort array of double values.
/// </summary>
/// <param name="values">Array of doubles which should be sorted.</param>
private void Sort( ref double [] values )
{
double tempValue;
for( int outLoop = 0; outLoop < values.Length; outLoop++ )
{
for( int inLoop = outLoop + 1; inLoop < values.Length; inLoop++ )
{
if( values[ outLoop ] > values[ inLoop ] )
{
tempValue = values[ outLoop ];
values[ outLoop ] = values[ inLoop ];
values[ inLoop ] = tempValue;
}
}
}
}
/// <summary>
/// Returns the median of the given numbers
/// </summary>
/// <param name="values">Array of double numbers</param>
/// <returns>Median</returns>
private double Median( double [] values )
{
// Exception for zero lenght of series.
if( values.Length == 0 )
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidMedianConditions);
}
// Sort array
Sort( ref values );
int position = values.Length / 2;
// if number of points is even
if( values.Length % 2 == 0 )
{
return ( values[position-1] + values[position] ) / 2.0;
}
else
{
return values[position];
}
}
/// <summary>
/// Calculates a Mean for a series of numbers.
/// </summary>
/// <param name="values">series with double numbers</param>
/// <returns>Returns Mean</returns>
private double Mean( double [] values )
{
// Exception for zero lenght of series.
if( values.Length == 0 )
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidMeanConditions);
}
// Find sum of values
double sum = 0;
foreach( double item in values )
{
sum += item;
}
// Calculate Mean
return sum / values.Length;
}
/// <summary>
/// Calculates a Variance for a series of numbers.
/// </summary>
/// <param name="values">double values</param>
/// <param name="sampleVariance">If variance is calculated from sample sum has to be divided by n-1.</param>
/// <returns>Variance</returns>
private double Variance( double [] values, bool sampleVariance )
{
// Exception for zero lenght of series.
if( values.Length < 1 )
{
throw new ArgumentException(SR.ExceptionStatisticalAnalysesInvalidVarianceConditions);
}
// Find sum of values
double sum = 0;
double mean = Mean( values );
foreach( double item in values )
{
sum += (item - mean) * (item - mean);
}
// Calculate Variance
if( sampleVariance )
{
return sum / ( values.Length - 1 );
}
else
{
return sum / values.Length;
}
}
#endregion // Statistical Parameters
# region Distributions
/// <summary>
/// Calculates the Percentage Points (probability) for the Student
/// t-distribution. The t-distribution is used in the hypothesis
/// testing of small sample data sets. Use this function in place
/// of a table of critical values for the t-distribution.
/// </summary>
/// <param name="tValue">The numeric value at which to evaluate the distribution.</param>
/// <param name="n">An integer indicating the number of degrees of freedom.</param>
/// <param name="oneTailed">Specifies the number of distribution tails to return.</param>
/// <returns>Returns the Percentage Points (probability) for the Student t-distribution.</returns>
private double StudentsDistribution( double tValue, int n, bool oneTailed )
{
// Validation
tValue = Math.Abs( tValue );
if( n > 300 )
{
n = 300;
}
if( n < 1 )
{
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesStudentsNegativeFreedomDegree);
}
double result = 1 - BetaIncomplete( n / 2.0, 0.5, n / (n + tValue * tValue) );
if( oneTailed )
return ( 1.0 - result ) / 2.0;
else
return 1.0 - result;
}
/// <summary>
/// Returns the standard normal cumulative distribution
/// function. The distribution has a mean of 0 (zero) and
/// a standard deviation of one. Use this function in place
/// of a table of standard normal curve areas.
/// </summary>
/// <param name="zValue">The value for which you want the distribution.</param>
/// <returns>Returns the standard normal cumulative distribution.</returns>
private double NormalDistribution( double zValue )
{
double [] a = {0.31938153,-0.356563782,1.781477937,-1.821255978,1.330274429};
double result;
if (zValue<-7.0)
{
result = NormalDistributionFunction(zValue)/Math.Sqrt(1.0+zValue*zValue);
}
else if (zValue>7.0)
{
result = 1.0 - NormalDistribution(-zValue);
}
else
{
result = 0.2316419;
result=1.0/(1+result*Math.Abs(zValue));
result=1-NormalDistributionFunction(zValue)*(result*(a[0]+result*(a[1]+result*(a[2]+result*(a[3]+result*a[4])))));
if (zValue<=0.0)
result=1.0-result;
}
return result;
}
private double FDistribution( double x, int freedom1, int freedom2 )
{
if (x < 0)
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesInvalidTValue);
if (freedom1 <= 0)
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesInvalidDegreeOfFreedom);
if (freedom2 <= 0)
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesInvalidDegreeOfFreedom);
if (x == 0)
return 1;
if (x == double.PositiveInfinity)
return 0;
return BetaIncomplete( freedom2 / 2.0, freedom1 / 2.0, freedom2 / ( freedom2 + freedom1 * x ) );
}
#endregion // Distributions
# region Inverse Distributions
/// <summary>
/// Calculates the t-value of the Student's t-distribution
/// as a function of the probability and the degrees of freedom.
/// </summary>
/// <param name="probability">The probability associated with the two-tailed Student's t-distribution.</param>
/// <param name="n">The number of degrees of freedom to characterize the distribution.</param>
/// <returns>Returns the t-value of the Student's t-distribution.</returns>
private double StudentsDistributionInverse( double probability, int n )
{
//Fix for boundary cases
if (probability == 0)
return double.PositiveInfinity;
else if (probability == 1)
return 0;
else if (probability < 0 || probability > 1)
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesInvalidProbabilityValue);
int step = 0;
return StudentsDistributionSearch( probability, n, step, 0.0, 100000.0 );
}
/// <summary>
/// Method for calculation of Inverse T Distribution (Binary tree)
/// solution for non linear equations
/// </summary>
/// <param name="probability">Probability value</param>
/// <param name="n">Degree of freedom</param>
/// <param name="step">Step for Numerical solution for non linear equations</param>
/// <param name="start">Start for numerical process</param>
/// <param name="end">End for numerical process</param>
/// <returns>Returns F ditribution inverse</returns>
private double StudentsDistributionSearch( double probability, int n, int step, double start, double end )
{
step++;
double mid = ( start + end ) / 2.0;
double result = StudentsDistribution( mid, n, false );
double resultX;
if( step > 100 )
{
return mid;
}
if( result <= probability )
{
resultX = StudentsDistributionSearch( probability, n, step, start, mid );
}
else
{
resultX = StudentsDistributionSearch( probability, n, step, mid, end );
}
return resultX;
}
/// <summary>
/// Returns the inverse of the standard normal cumulative distribution.
/// The distribution has a mean of zero and a standard deviation of one.
/// </summary>
/// <param name="probability">A probability corresponding to the normal distribution.</param>
/// <returns>Returns the inverse of the standard normal cumulative distribution.</returns>
private double NormalDistributionInverse( double probability )
{
// Validation
if( probability < 0.00001 || probability > 0.99999 )
{
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesNormalInvalidProbabilityValue);
}
double [] a = { 2.50662823884, -18.61500062529, 41.39119773534, -25.44106049637 };
double [] b = { -8.47351093090, 23.08336743743, -21.06224101826, 3.13082909833 };
double [] c = { 0.3374754822726147, 0.9761690190917186, 0.1607979714918209, 0.0276438810333863, 0.0038405729373609, 0.0003951896511919, 0.0000321767881768, 0.0000002888167364, 0.0000003960315187};
double x,r;
// Numerical Integration
x = probability - 0.5;
if ( Math.Abs(x) < 0.42 )
{
r = x * x;
r = x * ( ( ( a[3] * r + a[2] ) * r + a[1] ) * r + a[0] ) / ( ( ( ( b[3] * r + b[2] ) * r + b[1] ) * r + b[0] ) * r + 1.0 );
return( r );
}
r= probability;
if( x > 0.0 )
{
r = 1.0 - probability;
}
r = Math.Log( -Math.Log( r ) );
r = c[0] + r * ( c[1] + r * ( c[2] + r * ( c[3] + r * ( c[4] + r * ( c[5] + r * ( c[6] + r * ( c[7]+r * c[8] ) ) ) ) ) ) );
if( x < 0.0 )
{
r = -r;
}
return r;
}
/// <summary>
/// Calculates the inverse of the F probability distribution.
/// The F distribution can be used in an F-test that compares
/// the degree of variability in two data sets.
/// </summary>
/// <param name="probability">A probability associated with the F cumulative distribution.</param>
/// <param name="m">The numerator degrees of freedom.</param>
/// <param name="n">The denominator degrees of freedom.</param>
/// <returns>Returns the inverse of the F probability distribution.</returns>
private double FDistributionInverse( double probability, int m, int n )
{
//Fix for boundary cases
if (probability == 0)
return double.PositiveInfinity;
else if (probability == 1)
return 0;
else if (probability < 0 || probability > 1)
throw new ArgumentOutOfRangeException(SR.ExceptionStatisticalAnalysesInvalidProbabilityValue);
int step = 0;
return FDistributionSearch( probability, m, n, step, 0.0, 10000.0 );
}
/// <summary>
/// Method for calculation of Inverse F Distribution (Binary tree)
/// solution for non linear equations
/// </summary>
/// <param name="probability">Probability value</param>
/// <param name="m">Degree of freedom</param>
/// <param name="n">Degree of freedom</param>
/// <param name="step">Step for solution for non linear equations.</param>
/// <param name="start">Start for numerical process</param>
/// <param name="end">End for numerical process</param>
/// <returns>Returns F ditribution inverse</returns>
private double FDistributionSearch( double probability, int m, int n, int step, double start, double end )
{
step++;
double mid = ( start + end ) / 2.0;
double result = FDistribution( mid, m, n );
double resultX;
if( step > 30 )
{
return mid;
}
if( result <= probability )
{
resultX = FDistributionSearch( probability, m, n, step, start, mid );
}
else
{
resultX = FDistributionSearch( probability, m, n, step, mid, end );
}
return resultX;
}
#endregion // Inverse Distributions
}
}
|