public string writeModel(string outModelPath)
{
outPath = outModelPath;
using (System.IO.StreamWriter sw = new System.IO.StreamWriter(outPath))
{
sw.WriteLine(modelTypes.GLM.ToString());
sw.WriteLine(InTablePath);
sw.WriteLine(String.Join(",", IndependentFieldNames));
sw.WriteLine(String.Join(",", DependentFieldNames));
sw.WriteLine(String.Join(",", ClassFieldNames));
sw.WriteLine(SampleSize.ToString());
sw.WriteLine(NumberOfVariables.ToString());
sw.WriteLine(Iterations.ToString());
sw.WriteLine(DeltaC.ToString());
sw.WriteLine(LogLikelihood);
sw.WriteLine(LogLikelihoodratio);
sw.WriteLine(PValue.ToString());
sw.WriteLine(Deviance.ToString());
sw.WriteLine(ChiSquare.ToString());
sw.WriteLine(linkfunction.ToString());
sw.WriteLine(String.Join(" ", (from double d in Coefficients select d.ToString()).ToArray()));
sw.WriteLine(String.Join(" ", (from double d in StdError select d.ToString()).ToArray()));
sw.WriteLine(String.Join(" ", (from double d in waldTestValues select d.ToString()).ToArray()));
sw.WriteLine(String.Join(" ", (from double d in waldTestPValues select d.ToString()).ToArray()));
sw.WriteLine(String.Join(" ", (from double d in minValues select d.ToString()).ToArray()));
sw.WriteLine(String.Join(" ", (from double d in maxValues select d.ToString()).ToArray()));
sw.WriteLine(String.Join(" ", (from double d in sumValues select d.ToString()).ToArray()));
sw.Close();
}
return(outPath);
}
public void CanSampleSequence()
{
var n = new ChiSquare(1.0);
var ied = n.Samples();
ied.Take(5).ToArray();
}
public void getReport(double alpha)
{
Forms.RunningProcess.frmRunningProcessDialog rd = new Forms.RunningProcess.frmRunningProcessDialog(false);
rd.Text = "GLM Results";
rd.TopLevel = true;
rd.pgbProcess.Visible = false;
rd.FormBorderStyle = System.Windows.Forms.FormBorderStyle.Sizable;
rd.addMessage("Dependent field = " + DependentFieldNames[0]);
rd.addMessage("Independent fields = " + String.Join(", ", IndependentFieldNames));
rd.addMessage("Sample size = " + SampleSize.ToString());
rd.addMessage("Iteration = " + Iterations.ToString());
rd.addMessage("Delta Convergence " + DeltaC.ToString());
rd.addMessage("Chi-Sqr = " + ChiSquare.ToString() + " p-value = " + PValue.ToString());
rd.addMessage("Deviance = " + Deviance.ToString());
rd.addMessage("Log Likelihood = " + LogLikelihood.ToString());
rd.addMessage("Log Likelihood Ratio = " + LogLikelihoodratio.ToString() + "\n\nCoefficents and standard errors:\n");
rd.addMessage("Param: Intercept, " + String.Join(", ", IndependentFieldNames));
rd.addMessage("Coef: " + string.Join(", ", (from double d in Coefficients select d.ToString()).ToArray()));
rd.addMessage("STE: " + string.Join(", ", (from double d in StdError select d.ToString()).ToArray()) + "\n");
try
{
if (ModelHelper.chartingAvailable() && System.Windows.Forms.MessageBox.Show("Do you want to build distribution graphs?", "Graphs", System.Windows.Forms.MessageBoxButtons.YesNo) == System.Windows.Forms.DialogResult.Yes)
{
createRegChart();
}
}
catch
{
System.Windows.Forms.MessageBox.Show("Cannot create charts.");
}
rd.Show();
rd.enableClose();
}
public void SetDofFailsWithNonPositiveDoF(double dof)
{
{
var n = new ChiSquare(1.0);
n.DegreesOfFreedom = dof;
}
}
public List <string> GetDemographicsWhereThereIsAssociation()
{
List <string> demographicsWithAssociation = new List <string>();
List <int> FOT = new FunctionalOrganizationTypesServices().GetByCompany(Test.Company_Id).Select(fot => fot.Id).ToList();
//List<int> countriesId = new CountriesServices().GetCountriesByTest(Test);
CountriesServices countryService = new CountriesServices();
FunctionalOrganizationTypesServices FOTService = new FunctionalOrganizationTypesServices();
ChiSquare cs;
foreach (string demographic in DemographicNames.Keys) //recorro los demograficos
{
if (demographic == "FunctionalOrganizationType") //si el demografico actual es una estructura funcional
{
foreach (int type in FOT)
{
cs = new ChiSquare(Test, demographic, null, null, null, null, type, 0.05, country, state, region);
GetConclusion(demographicsWithAssociation, cs, demographic, FOTService.GetById(type).Name);
}
}
else
{
cs = new ChiSquare(Test, demographic, null, null, null, 0.05, country, state, region);
GetConclusion(demographicsWithAssociation, cs, demographic, "");
}
}
if (demographicsWithAssociation.Count > 0)
{
return(demographicsWithAssociation);
}
else
{
return(null);
}
}
/// <summary>
/// Creates an instance.
/// </summary>
/// <param name="degrees"> The degrees of freedom of the distribution, not less than one </param>
/// <param name="engine"> A uniform random number generator, not null </param>
public ChiSquareDistribution(double degrees, RandomEngine engine)
{
ArgChecker.isTrue(degrees >= 1, "Degrees of freedom must be greater than or equal to one");
ArgChecker.notNull(engine, "engine");
_chiSquare = new ChiSquare(degrees, engine);
_degrees = degrees;
}
/// <summary>
/// GOF test for one categorical variable with more than two levels.
///
/// Hypotheses are:
/// H_0 : actual distribution of each level = expected distribution of each level
/// H_1 : actual distribution of each level != expected distribution of each level
///
/// p-value = P(observed or more mismatch of expected and actual level distribution | H_0 is true)
///
/// Reject H_0 if p-value < alpha
/// </summary>
/// <param name="countOfEachLevel">The count of each level in the sample data for the categorical variable</param>
/// <param name="expectedPercentageOfEachLevel">The expected distribution / percentage of each level in the population for the categorical variable</param>
/// <param name="pValue">p-value which is P(observed or more extreme mismatch of expected and actual level distribution | H_0 is true</param>
/// <param name="significance_level">alpha</param>
/// <returns>True if H_0 is rejected; False if H_0 is failed to be rejected</returns>
public bool RejectH0(int[] observedCountInEachLevel, double[] expectedPercentageOfEachLevel, out double pValue, double significance_level = 0.05)
{
int sampleSize = 0;
int countOfLevels = observedCountInEachLevel.Length;
for (int i = 0; i < countOfLevels; ++i)
{
sampleSize += observedCountInEachLevel[i];
}
int[] expectedCountInEachLevel = new int[countOfLevels];
int r = sampleSize;
for (int i = 0; i < countOfLevels; ++i)
{
expectedCountInEachLevel[i] = (int)(expectedPercentageOfEachLevel[i] * sampleSize);
r -= expectedCountInEachLevel[i];
}
if (r > 0)
{
expectedCountInEachLevel[0] += r;
}
double ChiSq = 0;
for (int i = 0; i < countOfLevels; ++i)
{
ChiSq += System.Math.Pow(observedCountInEachLevel[i] - expectedCountInEachLevel[i], 2) / expectedCountInEachLevel[i];
}
pValue = 1 - ChiSquare.GetPercentile(ChiSq, countOfLevels - 1);
return(pValue < significance_level);
}
private Dictionary <string, double[]> GetSatisfiedAndNonSatisfiedByDemographic(string demographic, int?fot = null)
{
double EvaluationsByUbication = this.evaluationsByUbicationCount;// GetEvaluationsByUbication(country, state, region).Count();
Dictionary <string, double[]> data = new Dictionary <string, double[]>();
ChiSquare cs = new ChiSquare(Test, demographic, null, null, null, null, fot, 0.05, country, state, region);
Dictionary <string, double> satisfied = (Dictionary <string, double>)cs.DataSatisfaction["Satisfied"];
Dictionary <string, double> nonSatisfied = (Dictionary <string, double>)cs.DataSatisfaction["NoSatisfied"];
Dictionary <string, object> AvgAndMed = new Dictionary <string, object>();
Dictionary <string, double> average = new Dictionary <string, double>();
Dictionary <string, double> median = new Dictionary <string, double>();
List <string> keys = satisfied.Keys.ToList();
bool table = false;
if (keys.Count > 10)
{
if (demographic == "Location")
{
AvgAndMed = Test.GetAvgAndMedByLocations(null, null, null, false, country, state, region);
}
else if (demographic == "FunctionalOrganizationType")
{
AvgAndMed = Test.GetAvgAndMedByFOTypes(null, null, null, fot.Value, false, country, state, region);
}
else if (demographic == "AgeRange")
{
AvgAndMed = Test.GetAvgAndMedByAgeRanges(null, null, null, false, country, state, region);
}
else
{
AvgAndMed = Test.GetCategoryAvgAndMed(false, null, country, state, region);
}
average = (Dictionary <string, double>)AvgAndMed["Average"];
median = (Dictionary <string, double>)AvgAndMed["Median"];
table = true;
}
foreach (string key in keys)
{
double sat = satisfied[key];
double nonSat = nonSatisfied[key];
if (demographic == "Location")
{
EvaluationsByUbication = sat + nonSat;
}
double pctSat = sat * 100 / EvaluationsByUbication;
double pctNonSat = nonSat * 100 / EvaluationsByUbication;
if (table)
{
double avg = average[key];
double med = median[key];
data.Add(key, new double[] { sat, pctSat, nonSat, pctNonSat, avg, med });
}
else
{
data.Add(key, new double[] { sat, pctSat, nonSat, pctNonSat });
}
}
return(data);
}
public void ValidateDensity(
[Values(1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.5, 2.5, 2.5, 2.5, 2.5, 2.5, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity)] double dof,
[Values(0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity)] double x)
{
var n = new ChiSquare(dof);
Assert.AreEqual((Math.Pow(x, (dof / 2.0) - 1.0) * Math.Exp(-x / 2.0)) / (Math.Pow(2.0, dof / 2.0) * SpecialFunctions.Gamma(dof / 2.0)), n.Density(x));
}
public void ValidateDensityLn(
[Values(1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.5, 2.5, 2.5, 2.5, 2.5, 2.5, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity)] double dof,
[Values(0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity)] double x)
{
var n = new ChiSquare(dof);
Assert.AreEqual((-x / 2.0) + (((dof / 2.0) - 1.0) * Math.Log(x)) - ((dof / 2.0) * Math.Log(2)) - SpecialFunctions.GammaLn(dof / 2.0), n.DensityLn(x));
}
public void ValidateCumulativeDistribution(
[Values(1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.5, 2.5, 2.5, 2.5, 2.5, 2.5, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity)] double dof,
[Values(0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity)] double x)
{
var n = new ChiSquare(dof);
Assert.AreEqual(SpecialFunctions.GammaLowerIncomplete(dof / 2.0, x / 2.0) / SpecialFunctions.Gamma(dof / 2.0), n.CumulativeDistribution(x));
}
public ChiSquare GetSatisfiedAndNoSatisfiedDictionary(string demographic, int test_id, string type, double pValue, int?FO_id)
{
Test test = new TestsServices().GetById(test_id);
ChiSquare chiSquare = new ChiSquare(test, demographic, null, null, null, null, FO_id, pValue, null, null, null);
chiSquare.GetAssociation();
return(chiSquare);
}
public override void SaveParamsForReport()
{
ReportGenerator g = ReportGenerator.Instance;
g.ClearReflModelInfo();
int arrayconst = 2;
List <string> ginfo = new List <string>();
if (HoldsigmaCB)
{
ginfo.Add("The reflectivity curve was fit with a single roughness parameter\n");
}
else
{
ginfo.Add(String.Format("The reflectivity curve was fit with {0} roughness parameters\n", BoxCountTB + 1));
arrayconst = 3;
}
ginfo.Add("Percent Error in Q: " + QSpreadTB.ToString() + "\n");
ginfo.Add("Normalization Factor: " + NormalizationFactor.ToString() + "\n");
ginfo.Add("Low Q Offset: " + LowQOffset.ToString() + "\n");
ginfo.Add("High Q Offset: " + HighQOffset.ToString() + "\n");
ginfo.Add("Superphase SLD: " + SubphaseSLDTB.ToString() + "\n");
ginfo.Add("Subphase SLD: " + SubphaseSLD.ToString() + "\n");
ginfo.Add("Wavelength: " + WavelengthTB.ToString() + "\n");
ginfo.Add("Chi Square for reflectivity fit: " + ChiSquare.ToString() + "\n");
ginfo.Add(string.Format("The subphase roughness was: {0:#.### E-0} " + (char)0x00B1 + " {1:#.### E-0}\n", SubRoughTB, CovarArray[0]));
for (int i = 0; i < BoxCountTB; i++)
{
ginfo.Add((i + 1).ToString());
ginfo.Add(LengthArray[i].ToString("#.### E-0") + " " + (char)0x00B1 + " " + CovarArray[arrayconst * i + 1].ToString("#.### E-0"));
if (!m_bUseSLD)
{
ginfo.Add(RhoArray[i].ToString("#.### E-0") + " " + (char)0x00B1 + " " + CovarArray[arrayconst * i + 2].ToString("#.### E-0"));
}
else
{
ginfo.Add((RhoArray[i] * SubphaseSLDTB).ToString("#.### E-0") + " " + (char)0x00B1 + " " + (CovarArray[arrayconst * i + 2] * SubphaseSLDTB).ToString("#.### E-0"));
}
if (HoldsigmaCB)
{
ginfo.Add(SigmaArray[i].ToString("#.### E-0") + " " + (char)0x00B1 + " " + CovarArray[0].ToString("#.### E-0"));
}
else
{
ginfo.Add(SigmaArray[i].ToString("#.### E-0") + " " + (char)0x00B1 + " " + CovarArray[arrayconst * i + 3].ToString("#.### E-0"));
}
}
g.SetReflModelInfo = ginfo;
}
private static void CreateVectorMachines()
{
var kernel = new ChiSquare();
// Extract training parameters from the interface
double complexity = (double)1;
double tolerance = (double)0.01;
int cacheSize = (int)500;
SelectionStrategy strategy = SelectionStrategy.Sequential;
// Create the support vector machine learning algorithm
var teacher = new MulticlassSupportVectorLearning <IKernel>()
{
Kernel = kernel,
Learner = (param) =>
{
return(new SequentialMinimalOptimization <IKernel>()
{
Kernel = kernel,
Complexity = complexity,
Tolerance = tolerance,
CacheSize = cacheSize,
Strategy = strategy,
});
}
};
// Get the input and output data
double[][] inputs;
int[] outputs;
var inputsList = new List <double[]>();
var outputsList = new List <int>();
var i = 0;
foreach (var trainingImage in trainingImages.Keys)
{
var trainingFeature = trainingFeatures[trainingImage];
inputsList.Add(trainingFeature);
outputsList.Add(i);
i++;
}
inputs = inputsList.ToArray();
outputs = outputsList.ToArray();
ksvm = teacher.Learn(inputs, outputs);
double error = new ZeroOneLoss(outputs).Loss(ksvm.Decide(inputs));
Console.WriteLine("Error was {0}", error);
}
/// <summary>
/// Run example
/// </summary>
public void Run()
{
// 1. Get 10 random samples of f(x) = (x * x) / 2 using continuous uniform distribution on [-10, 10]
var uniform = new ContinuousUniform(-10, 10);
var result = SignalGenerator.Random(Function, uniform, 10);
Console.WriteLine(@" 1. Get 10 random samples of f(x) = (x * x) / 2 using continuous uniform distribution on [-10, 10]");
for (var i = 0; i < result.Length; i++)
{
Console.Write(result[i].ToString("N") + @" ");
}
Console.WriteLine();
Console.WriteLine();
// 2. Get 10 random samples of f(x) = (x * x) / 2 using Exponential(1) distribution and retrieve sample points
var exponential = new Exponential(1);
double[] samplePoints;
result = SignalGenerator.Random(Function, exponential, 10, out samplePoints);
Console.WriteLine(@"2. Get 10 random samples of f(x) = (x * x) / 2 using Exponential(1) distribution and retrieve sample points");
Console.Write(@"Points: ");
for (var i = 0; i < samplePoints.Length; i++)
{
Console.Write(samplePoints[i].ToString("N") + @" ");
}
Console.WriteLine();
Console.Write(@"Values: ");
for (var i = 0; i < result.Length; i++)
{
Console.Write(result[i].ToString("N") + @" ");
}
Console.WriteLine();
Console.WriteLine();
// 3. Get 10 random samples of f(x, y) = (x * y) / 2 using ChiSquare(10) distribution
var chiSquare = new ChiSquare(10);
result = SignalGenerator.Random(TwoDomainFunction, chiSquare, 10);
Console.WriteLine(@" 3. Get 10 random samples of f(x, y) = (x * y) / 2 using ChiSquare(10) distribution");
for (var i = 0; i < result.Length; i++)
{
Console.Write(result[i].ToString("N") + @" ");
}
Console.WriteLine();
}
public double GetSatisfiedCountPercentage()
{
ChiSquare cs = new ChiSquare(Test, "General", null, null, null, 0.05, country, state, region);
Dictionary <string, double> satisfied = (Dictionary <string, double>)cs.DataSatisfaction["Satisfied"];
if (satisfied.Count > 0)
{
int sat = (int)satisfied[Test.Name];
double pct = (double)(sat * 100) / evaluationsByUbicationCount; // GetEvaluationsByUbication(country, state, region).Count();
return(pct);
}
else
{
return(0);
}
}
public ChiSquare getChiSquare(int test_id, string Demographic, int?questionnaire_id, int?category_id, int?question_id, int?country_id, int?FO_id, double?pValue)
{
ChiSquare chiSquare;
double pvalue;
if (pValue.HasValue)
{
pvalue = pValue.Value;
}
else
{
pvalue = 0.05;
}
chiSquare = new ChiSquare(new TestsServices().GetById(test_id), Demographic, questionnaire_id, category_id, question_id, country_id, FO_id, pvalue, null, null, null);
chiSquare.GetAssociation();
return(chiSquare);
}
private void setKernalType(KernelType k)
{
switch (k)
{
case KernelType.Linear:
kernel = new Linear();
break;
case KernelType.Quadratic:
kernel = new Quadratic();
break;
case KernelType.Sigmoid:
kernel = new Sigmoid();
break;
case KernelType.Spline:
kernel = new Spline();
break;
case KernelType.ChiSquared:
kernel = new ChiSquare();
break;
case KernelType.Gaussian:
kernel = new Gaussian();
break;
case KernelType.Multiquadric:
kernel = new Multiquadric();
break;
case KernelType.InverseMultquadric:
kernel = new InverseMultiquadric();
break;
case KernelType.Laplacian:
kernel = new Laplacian();
break;
default:
break;
}
}
private void GetConclusion(List <string> demographicsWithAssociation, ChiSquare cs, string demographic, string name)
{
Dictionary <string, double> sat = (Dictionary <string, double>)cs.DataSatisfaction["Satisfied"];//diccionario de satisfechos y no satisfechos
if (sat.Count > 1)
{
cs.GetAssociation();//aqui realmente realiza lo de chicuadrado y trae los valores de:la tabla, de nosotros y la conclusion
}
if (cs.Association)
{
if (demographic == "FunctionalOrganizationType")
{
demographicsWithAssociation.Add(name);//agreega el nombre de la estructura funcional o el nombre del demografico
}
else
{
demographicsWithAssociation.Add(DemographicNames[demographic]);
}
}
}
public override string StochFit(double[] parampercs, int iterations)
{
double[] parameters = null;
string chosenchisquare = string.Empty;
MakeParameters(ref parameters, false);
double[] ParamArray = new double[1000 * parameters.Length];
double[] ChiSquareArray = new double[1000];
double[] CovarArray = new double[1000 * parameters.Length];
double[] locinfo = new double[9 * 1000];
int Size = 0;
InfoStruct = new BoxModelSettings();
SetInitStruct(ref InfoStruct, parampercs);
InfoStruct.Iterations = iterations;
NativeMethods.ConstrainedStochFit(InfoStruct, parameters, CovarArray, parameters.Length, locinfo, ParamArray, ChiSquareArray, ref Size);
//Not ideal, will always back up regardless of whether the new model is accepted or not
BackupArrays();
if (ModelChooser != null)
{
if (ModelChooser(ParamArray, ChiSquareArray, CovarArray, locinfo, Size, parameters.Length, InfoStruct))
{
}
}
UpdateProfile();
InfoStruct.Dispose();
return(ChiSquare.ToString("##.### E-0"));
}
public void CanSample()
{
var n = new ChiSquare(1.0);
n.Sample();
}
public void CanSampleStatic()
{
ChiSquare.Sample(new Random(), 2.0);
}
public void ValidateDensity(double dof, double x)
{
var n = new ChiSquare(dof);
Assert.AreEqual((Math.Pow(x, (dof / 2.0) - 1.0) * Math.Exp(-x / 2.0)) / (Math.Pow(2.0, dof / 2.0) * SpecialFunctions.Gamma(dof / 2.0)), n.Density(x));
}
public void ValidateMean(double dof)
{
var n = new ChiSquare(dof);
Assert.AreEqual(dof, n.Mean);
}
public void ValidateToString()
{
var n = new ChiSquare(1.0);
Assert.AreEqual("ChiSquare(DoF = 1)", n.ToString());
}
public void ValidateVariance(double dof)
{
var n = new ChiSquare(dof);
Assert.AreEqual<double>(2 * dof, n.Variance);
}
public void CanSetDoF(double dof)
{
var n = new ChiSquare(1.0);
n.DegreesOfFreedom = dof;
}
public void ValidateDensity(double dof, double x)
{
var n = new ChiSquare(dof);
Assert.AreEqual<double>((Math.Pow(x, dof / 2.0 - 1.0) * Math.Exp(-x / 2.0)) / (Math.Pow(2.0, dof / 2.0) * SpecialFunctions.Gamma(dof / 2.0)), n.Density(x));
}
public void ValidateDensityLn(double dof, double x)
{
var n = new ChiSquare(dof);
Assert.AreEqual<double>(-x / 2.0 + (dof / 2.0 - 1.0) * Math.Log(x) - (dof / 2.0) * Math.Log(2) - SpecialFunctions.GammaLn(dof / 2.0), n.DensityLn(x));
}
public void ValidateMedian(double dof)
{
var n = new ChiSquare(dof);
Assert.AreEqual<double>(dof - 2.0 / 3.0, n.Median);
}
public void ValidateMode(double dof)
{
var n = new ChiSquare(dof);
Assert.AreEqual<double>(dof - 2, n.Mode);
}
public void ValidateStdDev(double dof)
{
var n = new ChiSquare(dof);
Assert.AreEqual<double>(Math.Sqrt(n.Variance), n.StdDev);
}
public void ValidateCumulativeDistribution(double dof, double x)
{
var n = new ChiSquare(dof);
Assert.AreEqual(SpecialFunctions.GammaLowerIncomplete(dof / 2.0, x / 2.0) / SpecialFunctions.Gamma(dof / 2.0), n.CumulativeDistribution(x));
}
public void CanSample()
{
var n = new ChiSquare(1.0);
n.Sample();
}
public void ValidateToString()
{
var n = new ChiSquare(1.0);
Assert.AreEqual("ChiSquare(DoF = 1)", n.ToString());
}
public void ValidateCumulativeDistribution(
[Values(1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.5, 2.5, 2.5, 2.5, 2.5, 2.5, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity)] double dof,
[Values(0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity)] double x)
{
var n = new ChiSquare(dof);
Assert.AreEqual(SpecialFunctions.GammaLowerIncomplete(dof / 2.0, x / 2.0) / SpecialFunctions.Gamma(dof / 2.0), n.CumulativeDistribution(x));
}
public void ValidateDensityLn(
[Values(1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.5, 2.5, 2.5, 2.5, 2.5, 2.5, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity)] double dof,
[Values(0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity)] double x)
{
var n = new ChiSquare(dof);
Assert.AreEqual((-x / 2.0) + (((dof / 2.0) - 1.0) * Math.Log(x)) - ((dof / 2.0) * Math.Log(2)) - SpecialFunctions.GammaLn(dof / 2.0), n.DensityLn(x));
}
public void CanCreateChiSquare(double dof)
{
var n = new ChiSquare(dof);
Assert.AreEqual<double>(dof, n.DegreesOfFreedom);
}
public void CanSampleSequence()
{
var n = new ChiSquare(1.0);
var ied = n.Samples();
ied.Take(5).ToArray();
}
public void ValidateCumulativeDistribution(double dof, double x)
{
var n = new ChiSquare(dof);
Assert.AreEqual<double>(SpecialFunctions.GammaUpperIncomplete(dof / 2.0, x / 2.0) / SpecialFunctions.Gamma(dof / 2.0), n.CumulativeDistribution(x));
}
public void CanCreateChiSquare([Values(1.0, 3.0, Double.PositiveInfinity)] double dof)
{
var n = new ChiSquare(dof);
Assert.AreEqual(dof, n.DegreesOfFreedom);
}
public void SetDofFailsWithNonPositiveDoF(double dof)
{
{
var n = new ChiSquare(1.0);
n.DegreesOfFreedom = dof;
}
}
public void SetDofFailsWithNonPositiveDoF([Values(-1.0, -0.0, 0.0)] double dof)
{
var n = new ChiSquare(1.0);
Assert.Throws<ArgumentOutOfRangeException>(() => n.DegreesOfFreedom = dof);
}
public void ValidateVariance([Values(1.0, 2.0, 2.5, 3.0, Double.PositiveInfinity)] double dof)
{
var n = new ChiSquare(dof);
Assert.AreEqual(2 * dof, n.Variance);
}
public void ValidateMaximum()
{
var n = new ChiSquare(1.0);
Assert.AreEqual(Double.PositiveInfinity, n.Maximum);
}
public void ValidateStdDev([Values(1.0, 2.0, 2.5, 3.0, Double.PositiveInfinity)] double dof)
{
var n = new ChiSquare(dof);
Assert.AreEqual(Math.Sqrt(n.Variance), n.StdDev);
}
public void ValidateDensityLn(double dof, double x)
{
var n = new ChiSquare(dof);
Assert.AreEqual((-x / 2.0) + (((dof / 2.0) - 1.0) * Math.Log(x)) - ((dof / 2.0) * Math.Log(2)) - SpecialFunctions.GammaLn(dof / 2.0), n.DensityLn(x));
}
public void ValidateMode([Values(1.0, 2.0, 2.5, 3.0, Double.PositiveInfinity)] double dof)
{
var n = new ChiSquare(dof);
Assert.AreEqual(dof - 2, n.Mode);
}
public void FailSampleStatic()
{
Assert.Throws <ArgumentOutOfRangeException>(() => ChiSquare.Sample(new Random(), -1.0));
}
public void ValidateMedian([Values(1.0, 2.0, 2.5, 3.0, Double.PositiveInfinity)] double dof)
{
var n = new ChiSquare(dof);
Assert.AreEqual(dof - (2.0 / 3.0), n.Median);
}
public void ValidateMinimum()
{
var n = new ChiSquare(1.0);
Assert.AreEqual(0.0, n.Minimum);
}
public void ValidateMaximum()
{
var n = new ChiSquare(1.0);
Assert.AreEqual(Double.PositiveInfinity, n.Maximum);
}
public void CanCreateChiSquare(double dof)
{
var n = new ChiSquare(dof);
Assert.AreEqual(dof, n.DegreesOfFreedom);
}
public void ValidateDensity(
[Values(1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.5, 2.5, 2.5, 2.5, 2.5, 2.5, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity, Double.PositiveInfinity)] double dof,
[Values(0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity, 0.0, 0.1, 1.0, 5.5, 110.1, Double.PositiveInfinity)] double x)
{
var n = new ChiSquare(dof);
Assert.AreEqual((Math.Pow(x, (dof / 2.0) - 1.0) * Math.Exp(-x / 2.0)) / (Math.Pow(2.0, dof / 2.0) * SpecialFunctions.Gamma(dof / 2.0)), n.Density(x));
}
/// <summary>
/// Chi^2 independence test for categorical variables, var1 and var2
///
/// The hypotheses are:
/// H_0 : variable 1 is independent of variable 2
/// H_A : variable 1 and variable 2 are dependent
///
/// p-value = P(observed or more extreme events that favors H_A | H_0)
///
/// Now assuming H_0 is true, that is, the var1 and var2 are independent,
/// This implies the distribution of each level of var1 in each level of var2 should be the same
/// In other words, the expected distribution of each level of var1 in each level of var2 is given by distributionInEachLevel_var1
/// Now we can build a new contingency table containing the expected count corresponding to each level of both var1 and var2
///
/// Reject H_0 if p-value < alpha
/// </summary>
/// <param name="contingency_table">The contingency table in which each cell contains the counts of records in the sample data that matches the row (i.e. a var1 level) and col (i.e. a var2 level)</param>
/// <param name="pValue">p-value = P(observed or more extreme events that favors H_A | H_0)</param>
/// <param name="signficance_level">alpha</param>
/// <returns>True if H_0 is rejected; False if H_0 is failed to be rejected</returns>
public bool RejectH0(int[][] contingency_table, out double pValue, double signficance_level = 0.05)
{
int countOfLevels_var1 = contingency_table.Length;
int countOfLevels_var2 = contingency_table[0].Length;
int sampleSize = 0;
int[] countInEachLevel_var1 = new int[countOfLevels_var1];
for (int row = 0; row < countOfLevels_var1; ++row)
{
int countInLevel = 0;
for (int col = 0; col < countOfLevels_var2; ++col)
{
countInLevel += contingency_table[row][col];
}
countInEachLevel_var1[row] = countInLevel;
sampleSize += countInLevel;
}
double[] distributionInEachLevel_var1 = new double[countOfLevels_var1];
for (int row = 0; row < countOfLevels_var1; ++row)
{
distributionInEachLevel_var1[row] = (double)countInEachLevel_var1[row] / sampleSize;
}
int[] countInEachLevel_var2 = new int[countOfLevels_var2];
for (int col = 0; col < countOfLevels_var2; ++col)
{
int countInLevel = 0;
for (int row = 0; row < countOfLevels_var1; ++row)
{
countInLevel += contingency_table[row][col];
}
countInEachLevel_var2[col] = countInLevel;
}
//Now assuming H_0 is true, that is, the var1 and var2 are independent,
//This implies the distribution of each level of var1 in each level of var2 should be the same
//In other words, the expected distribution of each level of var1 in each level of var2 is given by distributionInEachLevel_var1
//Now we can build a new contingency table containing the expected count corresponding to each level of both var1 and var2
double[][] expected_contingency_table = new double[countOfLevels_var1][];
for (int row = 0; row < countOfLevels_var1; ++row)
{
expected_contingency_table[row] = new double[countOfLevels_var2];
for (int col = 0; col < countOfLevels_var2; ++col)
{
expected_contingency_table[row][col] = countInEachLevel_var2[col] * distributionInEachLevel_var1[row];
}
}
double ChiSq = 0;
for (int row = 0; row < countOfLevels_var1; ++row)
{
for (int col = 0; col < countOfLevels_var2; ++col)
{
ChiSq += System.Math.Pow(contingency_table[row][col] - expected_contingency_table[row][col], 2) / expected_contingency_table[row][col];
}
}
int df = (countOfLevels_var1 - 1) * (countOfLevels_var2 - 1);
pValue = 1 - ChiSquare.GetPercentile(ChiSq, df);
return(pValue < signficance_level);
}
public void ChiSquareCreateFailsWithBadParameters(double dof)
{
var n = new ChiSquare(dof);
}
