public double[] Correct(double[] observed)
{
if (observed == null || observed.Length != plexTable.Length)
{
throw new ArgumentException(string.Format("Argument observed should be a length {0} array, current is {1}", plexTable.Length, observed.Length));
}
var x = new double[observed.Length];
double rnorm = 0.0;
NonNegativeLeastSquaresCalc.NNLS(CloneTable(), observed.Length, plexTable[0].Length, observed, x, out rnorm, null, null, null);
return(x);
}
private SpeciesAbundanceInfo CalculateSpeciesAbundance(int maxLength, double[][] profileMatrix, PeakList <Peak> envelope, out double[] x)
{
double[] b = GetObservedIntensities(maxLength, envelope);
x = new double[3];
double rnorm;
NonNegativeLeastSquaresCalc.NNLS(profileMatrix, maxLength, 3, b, x, out rnorm, null, null, null);
var result = new SpeciesAbundanceInfo();
result.O16 = x[0];
result.O181 = x[1];
result.O182 = x[2];
return(result);
}
public static LinearRegressionRatioResult CalculateRatio(double[][] intensities)
{
double maxA = intensities[0].Max();
double maxB = intensities[1].Max();
if (maxA == 0.0)
{
return(new LinearRegressionRatioResult(20, 0.0));
}
if (maxB == 0.0)
{
return(new LinearRegressionRatioResult(0.05, 0.0));
}
if (intensities[0].Length == 1) // only one point
{
return(new LinearRegressionRatioResult(intensities[1][0] / intensities[0][0], 0.0));
}
var aList = new List <double>(intensities[0]);
aList.Add(0);
var a = new double[1][];
a[0] = aList.ToArray();
var bList = new List <double>(intensities[1]);
bList.Add(0);
double[] b = bList.ToArray();
var x = new double[1];
double rnorm = 0.0;
NonNegativeLeastSquaresCalc.NNLS(a, b.Length, 1, b, x, out rnorm, null, null, null);
var bPredicted = new double[aList.Count];
for (int i = 0; i < aList.Count; i++)
{
bPredicted[i] = x[0] * aList[i];
}
var result = new LinearRegressionRatioResult();
result.Ratio = x[0];
result.RSquare = StatisticsUtils.RSquare(bList.ToArray(), bPredicted);
result.PointCount = bList.Count;
result.TValue = StatisticsUtils.FCalculatedValueForLinearRegression(bList.ToArray(), bPredicted);
if (double.IsInfinity(result.TValue) || double.IsNaN(result.TValue) || result.TValue < 0)
{
result.TValue = 0;
}
result.PValue = StatisticsUtils.FProbabilityForLinearRegression(bList.Count, result.TValue);
return(result);
}
public static LinearRegressionRatioResult CalculateRatioDistance(double[][] intensities)
{
int len = intensities[0].Length;
double maxA = intensities[0].Max();
double maxB = intensities[1].Max();
if (maxA == 0.0)
{
return(new LinearRegressionRatioResult(20, 0.0));
}
if (maxB == 0.0)
{
return(new LinearRegressionRatioResult(0.05, 0.0));
}
if (len == 1) // only one point
{
return(new LinearRegressionRatioResult(intensities[1][0] / intensities[0][0], 0.0));
}
var a = new double[2][];
a[0] = new double[len];
a[1] = new double[len];
for (int i = 0; i < len; i++)
{
a[0][i] = intensities[0][i];
a[1][i] = 1;
}
double[] b = new List <double>(intensities[1]).ToArray();
var x = new double[2];
double rnorm = 0.0;
NonNegativeLeastSquaresCalc.NNLS(a, b.Length, 2, b, x, out rnorm, null, null, null);
var bPredicted = new double[len];
for (int i = 0; i < len; i++)
{
bPredicted[i] = x[0] * intensities[0][i] + x[1];
}
var result = new LinearRegressionRatioResult();
result.Ratio = x[0];
result.Distance = x[1];
result.RSquare = StatisticsUtils.RSquare(intensities[1], bPredicted);
result.PointCount = b.Length;
result.TValue = StatisticsUtils.FCalculatedValueForLinearRegression(intensities[1], bPredicted);
if (double.IsInfinity(result.TValue) || double.IsNaN(result.TValue) || result.TValue < 0)
{
result.TValue = 0;
}
result.PValue = StatisticsUtils.FProbabilityForLinearRegression(len, result.TValue);
return(result);
}
