0

I am trying to display a spectrum and it's upper and lower deviation. Therefor I have a Class, "SpectrumClass" where the cooridantes are stored in a DataTable (dtCoords).

The spectrum I am showing is an average of several other spectra. Whith those spectra I am calculating the standard-deviation as follows (all spectra have the same amount of datapoints):

DataTable dt = new DataTable();
        DataColumn columnX = new DataColumn("X");
        DataColumn columnY = new DataColumn("Y");

        dt.Columns.Add(columnX);
        dt.Columns.Add(columnY);

        SpectrumClass stdSpectrum = new SpectrumClass(0,"Standard deviation",dt);

        //Iterate through each Intensity value
        for (int i = 0; i < specs[0].dtCoords.Rows.Count; i++)
        {
            double rShift = 0;
            double IntensitySum = 0;

            //Calculate std-Points for each Intensity value
            foreach(SpectrumClass spec in specs)
            {
                IntensitySum += Convert.ToDouble(spec.dtCoords.Rows[i][1], System.Globalization.CultureInfo.InvariantCulture);
                rShift = Convert.ToDouble(spec.dtCoords.Rows[i ][0], System.Globalization.CultureInfo.InvariantCulture);
            }

            //Averge-value (Mittelwert)
            IntensitySum /= specs.Count;

            //Varianz
            double variance = 0;
            foreach (SpectrumClass spec in specs)
            {
                //Here IntensitySum is the average value
                variance += ((Convert.ToDouble(spec.dtCoords.Rows[i][1], System.Globalization.CultureInfo.InvariantCulture) - IntensitySum) * (Convert.ToDouble(spec.dtCoords.Rows[i][1], System.Globalization.CultureInfo.InvariantCulture) - IntensitySum));
            }

            //Std-Point
            double stdValue = variance /(specs.Count-1);

            stdValue = Math.Sqrt(stdValue);

            stdSpectrum.dtCoords.Rows.Add(rShift.ToString(System.Globalization.CultureInfo.InvariantCulture), stdValue.ToString(System.Globalization.CultureInfo.InvariantCulture));

        }

        return stdSpectrum;

Later when it comes to the visual part I create two new Series which should contain the upper and lower deviation. Here I am simply adding or aubtracting the Y-value for the according point of the average spectrum.

for (int i = dt.Rows.Count; i > 0;i--)
            {
                //Obere STD grenze
                if (expMeas.Count > 1)
                {

                    ExpChart.Series["STD+"].Points.AddXY(double.Parse(stdSpec.dtCoords.Rows[i - 1][0].ToString(), System.Globalization.CultureInfo.InvariantCulture)
                 , double.Parse(dt.Rows[i - 1][1].ToString(), System.Globalization.CultureInfo.InvariantCulture) + double.Parse(stdSpec.dtCoords.Rows[i - 1][1].ToString(), System.Globalization.CultureInfo.InvariantCulture));

                    ExpChart.Series["STD-"].Points.AddXY(double.Parse(stdSpec.dtCoords.Rows[i - 1][0].ToString(), System.Globalization.CultureInfo.InvariantCulture)
              , double.Parse(dt.Rows[i - 1][1].ToString(), System.Globalization.CultureInfo.InvariantCulture) - double.Parse(stdSpec.dtCoords.Rows[i - 1][1].ToString(), System.Globalization.CultureInfo.InvariantCulture));
                }

                ExpChart.Series["Spectrum"].Points.AddXY(double.Parse(stdSpec.dtCoords.Rows[i - 1][0].ToString(), System.Globalization.CultureInfo.InvariantCulture)
                 , double.Parse(dt.Rows[i - 1][1].ToString(), System.Globalization.CultureInfo.InvariantCulture));
            }

And here is my problem:

I think the standard-deviation spectra (in gray) should look the same?

I think the standard-deviation spectra (in gray) should look the same? I randomly checked standard-deviation points, they seem to be right. Could that be a mistake in one of my loops? I hope somebody experienced the same, beacuase checking throughover thousand points isn't fun

Thanks in advance.

4

1 回答 1

0

我想知道评论下是否有错误

//Obere STD grenze

在中间街区？

当你想调用 stdSpec 时，你是在调用 dt 吗？

double.Parse(dt.Rows[i - 1][0].ToString()

double.Parse(stdSpec.dtCoords.Rows[i - 1][0]

于 2013-06-04T18:21:21.687 回答