public void MeanTest()
{
double[] DataPoints = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0 };
DescriptiveStats statsCalc = new DescriptiveStats();
var result = statsCalc.Mean(DataPoints);
Assert.AreEqual(3.5, result);
}
public void SkewnessTest()
{
double[] DataPoints = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0 };
DescriptiveStats statsCalc = new DescriptiveStats();
var result = statsCalc.Skewness(DataPoints);
Assert.AreEqual(0.0, result);
}
public void StdDevTest()
{
double[] DataPoints = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0 };
DescriptiveStats statsCalc = new DescriptiveStats();
var result = statsCalc.StdDev(DataPoints);
Assert.AreEqual(1.8708286933869707, result);
}
public void MeanDeviationTest()
{
double[] DataPoints = { 12.0, 24.0, 41.0, 51.0, 67.0, 67.0, 85.0, 99.0 };
DescriptiveStats statsCalc = new DescriptiveStats();
var result = statsCalc.MeanDeviation(DataPoints);
Assert.AreEqual(23.75, result);
}
public void ZScoresTest()
{
double[] DataPoints = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0 };
double[] scores = { -1.3363062095621219, -0.449056948958286, 0.05601104036639116, 0.5284884801832448, 1.0840732389765304, 1.926539846563328 };
DescriptiveStats statsCalc = new DescriptiveStats();
var result = statsCalc.ZScore(DataPoints);
CollectionAssert.AreEqual(scores, result);
}
public void PopCorrelationTest()
{
double[] DataPointsX = { 1, 2, 3, 4, 5 };
double[] DataPointsY = { 11, 22, 34, 43, 56 };
DescriptiveStats statsCalc = new DescriptiveStats();
var result = statsCalc.PopCoefficient(DataPointsX, DataPointsY);
Assert.AreEqual(0.7991248857909374, result);
}
public void SamCorrelationTest()
{
double[] DataPointsX = { 3.0, 3.0, 6.0 };
double[] DataPointsY = { 2.0, 3.0, 4.0 };
DescriptiveStats statsCalc = new DescriptiveStats();
var result = statsCalc.SampleCoefficient(DataPointsX, DataPointsY);
Assert.AreEqual(0.8660254037844387, result);
}
/// <summary>
/// a hacked mess to show box-wisker plot for outlier manipulation
/// </summary>
/// <param name="iv">double array of independent variable values</param>
/// <param name="tags">string array of what should be date/time tags</param>
/// <returns>a graph pane to display in the master</returns>
private GraphPane addPlotBW(double[] iv, string[] tags)
{
//http://www.sharpstatistics.co.uk/index.php?option=com_content&view=article&id=12&Itemid=13
//and hacked by mog 4/1/11
//For each set of data for the boxplot calculate the median and the inter quartile (IQR) range
//which is just the value of the 75th percentile minus the 25th percentile. The median is where
//the horizontal bar goes. Using the 25th and 75th percentile a HiLowBarIten can be set which is
//just a bar where the top and base are specified.
//An error bar that has upper and lower values of the 25th percentile minus 1.5 times the IQR and
//the 75th percentile plus 1.5 times the IQR is set to give the whiskers. As with the HiLowItem
//barList is a PointPairList with the high and low values.
//All that is left to do is add any points that are above or below the end of the whiskers.
SortedList <string, double> datevalue = getlist(tags, iv);
GraphPane gp = new GraphPane();
// "Median designated within box.\n" +
// "Whiskers are +- 1.5 * IQR (IQR = 75th percentile - 25th percentile)\n" +
// "Points above/below whiskers are designated as outliers.";
//median of each array
PointPairList medians = new PointPairList();
//75th and 25th percentile, defines the box
PointPairList hiLowList = new PointPairList();
//+/- 1.5*Interquartile range, extentent of wiskers
PointPairList barList = new PointPairList();
//outliers
PointPairList outs = new PointPairList();
//Add the values
DescriptiveStats ds = new DescriptiveStats();
ds.getStats(iv);
double median = ds.Median;
medians.Add(0, median);
double hivalue = percentile(iv, 75);
double lovalue = percentile(iv, 25);
hiLowList.Add(0, hivalue, lovalue);
double iqr = 1.5 * (hivalue - lovalue);
double upperLimit = hivalue + iqr;
double lowerLimit = lovalue - iqr;
//The wiskers must end on an actual data point
double wiskerlo = ValueNearestButGreater(iv, lowerLimit);
double wiskerhi = ValueNearestButLess(iv, upperLimit);
barList.Add(0, wiskerlo, wiskerhi);
var upperouts = (from kv in datevalue
where kv.Value > upperLimit
select kv);
foreach (var v in upperouts)
{
outs.Add(0, v.Value, v.Key);
}
var lowerouts = (from kv in datevalue
where kv.Value < lowerLimit
select kv);
foreach (var v in lowerouts)
{
outs.Add(0, v.Value, v.Key);
}
//Plot the items, first the median values
CurveItem meadian = gp.AddCurve("", medians, Color.Black, SymbolType.HDash);
LineItem myLine = (LineItem)meadian;
myLine.Line.IsVisible = false;
myLine.Symbol.Fill.Type = FillType.Solid;
//Box
HiLowBarItem myCurve = gp.AddHiLowBar("", hiLowList, Color.Black);
myCurve.Bar.Fill.Type = FillType.None;
//Wiskers
ErrorBarItem myerror = gp.AddErrorBar("", barList, Color.Black);
//Outliers
CurveItem upper = gp.AddCurve(_dt.Columns[intSelectedcol].ColumnName + " outliers", outs, Color.Green, SymbolType.Circle);
LineItem bLine = (LineItem)upper;
bLine.Line.IsVisible = false;
gp.YAxis.Title.Text = _dt.Columns[intSelectedcol].ColumnName;
gp.BarSettings.Type = BarType.Overlay;
gp.XAxis.IsVisible = false;
gp.Legend.IsVisible = true;
gp.Tag = "BWPlot";
gp.Title.Text = "BoxWhisker Plot";
return(gp);
}
/// <summary>
/// shows some stats for the selected variable
/// </summary>
private void showColInfo()
{
string varName = _dt.Columns[_selectedcol].Caption;
var count = (from DataRow r in _dt.Rows select r[_selectedcol]).Count();
var max = (from DataRow r in _dt.Rows select r[_selectedcol]).Max();
var min = (from DataRow r in _dt.Rows select r[_selectedcol]).Min();
var avg = (from DataRow r in _dt.Rows select r[_selectedcol]).Average(r => (double)r);
var unique = (from DataRow r in _dt.Rows select r[_selectedcol]).Distinct().Count();
var zerocount = (from DataRow r in _dt.Rows where (double)r[_selectedcol] == 0 select r[_selectedcol]).Count();
double dmax = Convert.ToDouble(max);
double dmin = Convert.ToDouble(min);
AndersonDarlingNormality adtest = new AndersonDarlingNormality();
double[] vals = (from DataRow r in _dt.Rows select(double) r[_selectedcol]).ToArray <double>();
adtest.getADstat(vals);
double norm = adtest.ADStat;
double pvnorm = adtest.ADStatPval;
DescriptiveStats ds = new DescriptiveStats();
ds.getStats(vals);
double median = ds.Median; //nv.Median;
double range = ds.Range; // nv.Range;
double mean = ds.Mean; // nv.Mean;
double stddev = ds.StdDev; // nv.StandardDeviation;
double variance = ds.Variance; // nv.Variance;
double kurtosis = ds.Kurtosis; //nv.Kurtosis;
double skewness = ds.Kurtosis; //nv.Skewness;
listView1.View = View.Details;
ListViewItem lvi;
if (listView1.Items.Count > 0)
{
lvi = new ListViewItem("*****");
lvi.SubItems.Add("");
listView1.Items.Add(lvi);
}
lvi = new ListViewItem("Variable Name");
lvi.SubItems.Add(varName);
listView1.Items.Add(lvi);
lvi = new ListViewItem("Row Count");
lvi.SubItems.Add(count.ToString());
listView1.Items.Add(lvi);
lvi = new ListViewItem("Maximum Value");
lvi.SubItems.Add(dmax.ToString("n2"));
listView1.Items.Add(lvi);
lvi = new ListViewItem("Minimum Value");
lvi.SubItems.Add(dmin.ToString("n2"));
listView1.Items.Add(lvi);
lvi = new ListViewItem("Average Value");
lvi.SubItems.Add(avg.ToString("n2"));
listView1.Items.Add(lvi);
lvi = new ListViewItem("Unique Values");
lvi.SubItems.Add(unique.ToString());
listView1.Items.Add(lvi);
lvi = new ListViewItem("Zero Count");
lvi.SubItems.Add(zerocount.ToString());
listView1.Items.Add(lvi);
lvi = new ListViewItem("Median Value");
lvi.SubItems.Add(median.ToString("n3"));
listView1.Items.Add(lvi);
lvi = new ListViewItem("Data Range");
lvi.SubItems.Add(range.ToString("n3"));
listView1.Items.Add(lvi);
lvi = new ListViewItem("");
lvi.SubItems.Add("");
listView1.Items.Add(lvi);
lvi = new ListViewItem("AD Statistic");
lvi.SubItems.Add(_regressionadstat.ToString("n4"));
//lvi.SubItems.Add(norm.ToString("n3"));
listView1.Items.Add(lvi);
lvi = new ListViewItem("AD Stat P-Value");
lvi.SubItems.Add(_regressionadpval.ToString("n4"));
//lvi.SubItems.Add(pvnorm.ToString("n3"));
listView1.Items.Add(lvi);
lvi = new ListViewItem("Mean Value");
lvi.SubItems.Add(mean.ToString("n3"));
listView1.Items.Add(lvi);
lvi = new ListViewItem("Standard Deviation");
lvi.SubItems.Add(stddev.ToString("n3"));
listView1.Items.Add(lvi);
lvi = new ListViewItem("Variance");
lvi.SubItems.Add(variance.ToString("n3"));
listView1.Items.Add(lvi);
lvi = new ListViewItem("Kurtosis");
lvi.SubItems.Add(kurtosis.ToString("n3"));
listView1.Items.Add(lvi);
lvi = new ListViewItem("Skewness");
lvi.SubItems.Add(skewness.ToString("n3"));
listView1.Items.Add(lvi);
//magic numbers for widths: -1 set to max characters in subitems, -2 == autosize
listView1.Columns.Add("Data", -1, HorizontalAlignment.Right);
listView1.Columns.Add("Value", -2, HorizontalAlignment.Left);
}
