protected virtual void DeregisterScatterPlotDataRowEvents(ScatterPlotDataRow row) {
row.Points.ItemsAdded -= new CollectionItemsChangedEventHandler<IndexedItem<Point2D<double>>>(Points_ItemsAdded);
row.Points.ItemsRemoved -= new CollectionItemsChangedEventHandler<IndexedItem<Point2D<double>>>(Points_ItemsRemoved);
row.Points.ItemsReplaced -= new CollectionItemsChangedEventHandler<IndexedItem<Point2D<double>>>(Points_ItemsReplaced);
row.Points.CollectionReset -= new CollectionItemsChangedEventHandler<IndexedItem<Point2D<double>>>(Points_CollectionReset);
pointsRowsTable.Remove(row.Points);
row.VisualPropertiesChanged -= new EventHandler(Row_VisualPropertiesChanged);
row.NameChanged -= new EventHandler(Row_NameChanged);
}
private void ConfigureSeries(Series series, ScatterPlotDataRow row)
{
series.BorderWidth = 1;
series.BorderDashStyle = ChartDashStyle.Solid;
series.BorderColor = Color.Empty;
if (row.VisualProperties.Color != Color.Empty)
{
series.Color = row.VisualProperties.Color;
}
else
{
series.Color = Color.Empty;
}
series.IsVisibleInLegend = row.VisualProperties.IsVisibleInLegend;
series.ChartType = SeriesChartType.FastPoint;
series.MarkerSize = row.VisualProperties.PointSize;
series.MarkerStyle = ConvertPointStyle(row.VisualProperties.PointStyle);
series.XAxisType = AxisType.Primary;
series.YAxisType = AxisType.Primary;
if (row.VisualProperties.DisplayName.Trim() != String.Empty)
{
series.LegendText = row.VisualProperties.DisplayName;
}
else
{
series.LegendText = row.Name;
}
string xAxisTitle = string.IsNullOrEmpty(Content.VisualProperties.XAxisTitle)
? "X"
: Content.VisualProperties.XAxisTitle;
string yAxisTitle = string.IsNullOrEmpty(Content.VisualProperties.YAxisTitle)
? "Y"
: Content.VisualProperties.YAxisTitle;
series.ToolTip =
series.LegendText + Environment.NewLine +
xAxisTitle + " = " + "#VALX," + Environment.NewLine +
yAxisTitle + " = " + "#VAL";
}
public ScatterPlot CreateScatterPlot(string variableNameX, string variableNameY)
{
ScatterPlot scatterPlot = new ScatterPlot();
IList <double> xValues = PreprocessingData.GetValues <double>(PreprocessingData.GetColumnIndex(variableNameX));
IList <double> yValues = PreprocessingData.GetValues <double>(PreprocessingData.GetColumnIndex(variableNameY));
List <Point2D <double> > points = new List <Point2D <double> >();
for (int i = 0; i < xValues.Count; i++)
{
Point2D <double> point = new Point2D <double>(xValues[i], yValues[i]);
points.Add(point);
}
ScatterPlotDataRow scdr = new ScatterPlotDataRow(variableNameX + " - " + variableNameY, "", points);
scatterPlot.Rows.Add(scdr);
return(scatterPlot);
}
private void Row_VisualPropertiesChanged(object sender, EventArgs e)
{
if (InvokeRequired)
{
Invoke(new EventHandler(Row_VisualPropertiesChanged), sender, e);
}
else
{
ScatterPlotDataRow row = (ScatterPlotDataRow)sender;
Series series = chart.Series[row.Name];
Series regressionSeries = seriesToRegressionSeriesTable[series];
series.Points.Clear();
regressionSeries.Points.Clear();
ConfigureSeries(series, regressionSeries, row);
FillSeriesWithRowValues(series, row);
FillRegressionSeries(regressionSeries, row);
RecalculateMinMaxPointValues();
RecalculateAxesScale(chart.ChartAreas[0]);
UpdateRegressionSeriesColors();
}
}
private void Points_CollectionReset(object sender, CollectionItemsChangedEventArgs <IndexedItem <Point2D <double> > > e)
{
if (InvokeRequired)
{
Invoke(new CollectionItemsChangedEventHandler <IndexedItem <Point2D <double> > >(Points_CollectionReset), sender, e);
}
else
{
ScatterPlotDataRow row = null;
pointsRowsTable.TryGetValue((IObservableList <Point2D <double> >)sender, out row);
if (row != null)
{
Series rowSeries = chart.Series[row.Name];
if (!invisibleSeries.Contains(rowSeries))
{
rowSeries.Points.Clear();
FillSeriesWithRowValues(rowSeries, row);
RecalculateAxesScale(chart.ChartAreas[0]);
UpdateYCursorInterval();
}
}
}
}
protected static bool Fitting(ScatterPlotDataRow row, out double[] coefficients)
{
if (!IsValidRegressionData(row))
{
coefficients = new double[0];
return(false);
}
var xs = row.Points.Select(p => p.X).ToList();
var ys = row.Points.Select(p => p.Y).ToList();
// Input transformations
double[,] matrix;
int nRows;
switch (row.VisualProperties.RegressionType)
{
case ScatterPlotDataRowVisualProperties.ScatterPlotDataRowRegressionType.Linear:
matrix = CreateMatrix(out nRows, ys, xs);
break;
case ScatterPlotDataRowVisualProperties.ScatterPlotDataRowRegressionType.Polynomial:
var xss = Enumerable.Range(1, row.VisualProperties.PolynomialRegressionOrder)
.Select(o => xs.Select(x => Math.Pow(x, o)).ToList())
.Reverse(); // higher order first
matrix = CreateMatrix(out nRows, ys, xss.ToArray());
break;
case ScatterPlotDataRowVisualProperties.ScatterPlotDataRowRegressionType.Exponential:
matrix = CreateMatrix(out nRows, ys.Select(y => Math.Log(y)).ToList(), xs);
break;
case ScatterPlotDataRowVisualProperties.ScatterPlotDataRowRegressionType.Power:
matrix = CreateMatrix(out nRows, ys.Select(y => Math.Log(y)).ToList(), xs.Select(x => Math.Log(x)).ToList());
break;
case ScatterPlotDataRowVisualProperties.ScatterPlotDataRowRegressionType.Logarithmic:
matrix = CreateMatrix(out nRows, ys, xs.Select(x => Math.Log(x)).ToList());
break;
default:
throw new ArgumentException("Unknown RegressionType: " + row.VisualProperties.RegressionType);
}
// Linear fitting
bool success = LinearFitting(matrix, nRows, out coefficients);
if (!success)
{
return(false);
}
// Output transformation
switch (row.VisualProperties.RegressionType)
{
case ScatterPlotDataRowVisualProperties.ScatterPlotDataRowRegressionType.Exponential:
case ScatterPlotDataRowVisualProperties.ScatterPlotDataRowRegressionType.Power:
coefficients[1] = Math.Exp(coefficients[1]);
break;
}
return(true);
}
private void SetUpResults(IReadOnlyList <int> allIndices)
{
if (Results == null)
{
return;
}
var results = Results;
dataRowIndices = new Dictionary <string, IList <int> >();
var problemData = Problem.ProblemData;
if (!results.ContainsKey(IterationResultName))
{
results.Add(new Result(IterationResultName, new IntValue(0)));
}
if (!results.ContainsKey(ErrorResultName))
{
results.Add(new Result(ErrorResultName, new DoubleValue(0)));
}
if (!results.ContainsKey(ScatterPlotResultName))
{
results.Add(new Result(ScatterPlotResultName, "Plot of the projected data", new ScatterPlot(DataResultName, "")));
}
if (!results.ContainsKey(DataResultName))
{
results.Add(new Result(DataResultName, "Projected Data", new DoubleMatrix()));
}
if (!results.ContainsKey(ErrorPlotResultName))
{
var errortable = new DataTable(ErrorPlotResultName, "Development of errors during gradient descent")
{
VisualProperties =
{
XAxisTitle = "UpdateIntervall",
YAxisTitle = "Error",
YAxisLogScale = true
}
};
errortable.Rows.Add(new DataRow("Errors"));
errortable.Rows["Errors"].VisualProperties.StartIndexZero = true;
results.Add(new Result(ErrorPlotResultName, errortable));
}
//color datapoints acording to classes variable (be it double, datetime or string)
if (!problemData.Dataset.VariableNames.Contains(ClassesName))
{
dataRowIndices.Add("Training", problemData.TrainingIndices.ToList());
dataRowIndices.Add("Test", problemData.TestIndices.ToList());
return;
}
var classificationData = problemData as ClassificationProblemData;
if (classificationData != null && classificationData.TargetVariable.Equals(ClassesName))
{
var classNames = classificationData.ClassValues.Zip(classificationData.ClassNames, (v, n) => new { v, n }).ToDictionary(x => x.v, x => x.n);
var classes = classificationData.Dataset.GetDoubleValues(classificationData.TargetVariable, allIndices).Select(v => classNames[v]).ToArray();
for (var i = 0; i < classes.Length; i++)
{
if (!dataRowIndices.ContainsKey(classes[i]))
{
dataRowIndices.Add(classes[i], new List <int>());
}
dataRowIndices[classes[i]].Add(i);
}
}
else if (((Dataset)problemData.Dataset).VariableHasType <string>(ClassesName))
{
var classes = problemData.Dataset.GetStringValues(ClassesName, allIndices).ToArray();
for (var i = 0; i < classes.Length; i++)
{
if (!dataRowIndices.ContainsKey(classes[i]))
{
dataRowIndices.Add(classes[i], new List <int>());
}
dataRowIndices[classes[i]].Add(i);
}
}
else if (((Dataset)problemData.Dataset).VariableHasType <double>(ClassesName))
{
var clusterdata = new Dataset(problemData.Dataset.DoubleVariables, problemData.Dataset.DoubleVariables.Select(v => problemData.Dataset.GetDoubleValues(v, allIndices).ToList()));
const int contours = 8;
Dictionary <int, string> contourMap;
IClusteringModel clusterModel;
double[][] borders;
CreateClusters(clusterdata, ClassesName, contours, out clusterModel, out contourMap, out borders);
var contourorder = borders.Select((x, i) => new { x, i }).OrderBy(x => x.x[0]).Select(x => x.i).ToArray();
for (var i = 0; i < contours; i++)
{
var c = contourorder[i];
var contourname = contourMap[c];
dataRowIndices.Add(contourname, new List <int>());
var row = new ScatterPlotDataRow(contourname, "", new List <Point2D <double> >())
{
VisualProperties = { Color = GetHeatMapColor(i, contours), PointSize = 8 }
};
((ScatterPlot)results[ScatterPlotResultName].Value).Rows.Add(row);
}
var allClusters = clusterModel.GetClusterValues(clusterdata, Enumerable.Range(0, clusterdata.Rows)).ToArray();
for (var i = 0; i < clusterdata.Rows; i++)
{
dataRowIndices[contourMap[allClusters[i] - 1]].Add(i);
}
}
else if (((Dataset)problemData.Dataset).VariableHasType <DateTime>(ClassesName))
{
var clusterdata = new Dataset(problemData.Dataset.DateTimeVariables, problemData.Dataset.DateTimeVariables.Select(v => problemData.Dataset.GetDoubleValues(v, allIndices).ToList()));
const int contours = 8;
Dictionary <int, string> contourMap;
IClusteringModel clusterModel;
double[][] borders;
CreateClusters(clusterdata, ClassesName, contours, out clusterModel, out contourMap, out borders);
var contourorder = borders.Select((x, i) => new { x, i }).OrderBy(x => x.x[0]).Select(x => x.i).ToArray();
for (var i = 0; i < contours; i++)
{
var c = contourorder[i];
var contourname = contourMap[c];
dataRowIndices.Add(contourname, new List <int>());
var row = new ScatterPlotDataRow(contourname, "", new List <Point2D <double> >())
{
VisualProperties = { Color = GetHeatMapColor(i, contours), PointSize = 8 }
};
row.VisualProperties.PointSize = 8;
((ScatterPlot)results[ScatterPlotResultName].Value).Rows.Add(row);
}
var allClusters = clusterModel.GetClusterValues(clusterdata, Enumerable.Range(0, clusterdata.Rows)).ToArray();
for (var i = 0; i < clusterdata.Rows; i++)
{
dataRowIndices[contourMap[allClusters[i] - 1]].Add(i);
}
}
else
{
dataRowIndices.Add("Training", problemData.TrainingIndices.ToList());
dataRowIndices.Add("Test", problemData.TestIndices.ToList());
}
}
public override IOperation Apply()
{
var operation = base.Apply();
var paretoFront = TrainingBestSolutionsParameter.ActualValue;
IResult result;
ScatterPlot qualityToTreeSize;
if (!ResultCollection.TryGetValue("Pareto Front Analysis", out result))
{
qualityToTreeSize = new ScatterPlot("Quality vs Tree Size", "");
qualityToTreeSize.VisualProperties.XAxisMinimumAuto = false;
qualityToTreeSize.VisualProperties.XAxisMaximumAuto = false;
qualityToTreeSize.VisualProperties.YAxisMinimumAuto = false;
qualityToTreeSize.VisualProperties.YAxisMaximumAuto = false;
qualityToTreeSize.VisualProperties.XAxisMinimumFixedValue = 0;
qualityToTreeSize.VisualProperties.XAxisMaximumFixedValue = MaximumSymbolicExpressionTreeLengthParameter.ActualValue.Value;
qualityToTreeSize.VisualProperties.YAxisMinimumFixedValue = 0;
qualityToTreeSize.VisualProperties.YAxisMaximumFixedValue = 2;
ResultCollection.Add(new Result("Pareto Front Analysis", qualityToTreeSize));
}
else
{
qualityToTreeSize = (ScatterPlot)result.Value;
}
int previousTreeLength = -1;
var sizeParetoFront = new LinkedList <ISymbolicRegressionSolution>();
foreach (var solution in paretoFront.OrderBy(s => s.Model.SymbolicExpressionTree.Length))
{
int treeLength = solution.Model.SymbolicExpressionTree.Length;
if (!sizeParetoFront.Any())
{
sizeParetoFront.AddLast(solution);
}
if (solution.TrainingNormalizedMeanSquaredError < sizeParetoFront.Last.Value.TrainingNormalizedMeanSquaredError)
{
if (treeLength == previousTreeLength)
{
sizeParetoFront.RemoveLast();
}
sizeParetoFront.AddLast(solution);
}
previousTreeLength = treeLength;
}
qualityToTreeSize.Rows.Clear();
var trainingRow = new ScatterPlotDataRow("Training NMSE", "", sizeParetoFront.Select(x => new Point2D <double>(x.Model.SymbolicExpressionTree.Length, x.TrainingNormalizedMeanSquaredError, x)));
trainingRow.VisualProperties.PointSize = 8;
qualityToTreeSize.Rows.Add(trainingRow);
if (AnalyzeTestError)
{
var testRow = new ScatterPlotDataRow("Test NMSE", "",
sizeParetoFront.Select(x => new Point2D <double>(x.Model.SymbolicExpressionTree.Length, x.TestNormalizedMeanSquaredError, x)));
testRow.VisualProperties.PointSize = 8;
qualityToTreeSize.Rows.Add(testRow);
}
var validationPartition = ValidationPartitionParameter.ActualValue;
if (validationPartition.Size != 0)
{
var problemData = ProblemDataParameter.ActualValue;
var validationIndizes = Enumerable.Range(validationPartition.Start, validationPartition.Size).ToList();
var targetValues = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, validationIndizes).ToList();
OnlineCalculatorError error;
var validationRow = new ScatterPlotDataRow("Validation NMSE", "",
sizeParetoFront.Select(x => new Point2D <double>(x.Model.SymbolicExpressionTree.Length,
OnlineNormalizedMeanSquaredErrorCalculator.Calculate(targetValues, x.GetEstimatedValues(validationIndizes), out error))));
validationRow.VisualProperties.PointSize = 7;
qualityToTreeSize.Rows.Add(validationRow);
}
return(operation);
}
public static ScatterPlot CreateScatterPlot(IFilteredPreprocessingData preprocessingData, string variableNameX, string variableNameY, string variableNameGroup = "-", LegendOrder legendOrder = LegendOrder.Alphabetically)
{
ScatterPlot scatterPlot = new ScatterPlot();
IList <double> xValues = preprocessingData.GetValues <double>(preprocessingData.GetColumnIndex(variableNameX));
IList <double> yValues = preprocessingData.GetValues <double>(preprocessingData.GetColumnIndex(variableNameY));
var points = xValues.Zip(yValues, (x, y) => new Point2D <double>(x, y)).ToList();
var validPoints = points.Where(p => !double.IsNaN(p.X) && !double.IsNaN(p.Y) && !double.IsInfinity(p.X) && !double.IsInfinity(p.Y)).ToList();
if (validPoints.Any())
{
try {
double axisMin, axisMax, axisInterval;
ChartUtil.CalculateOptimalAxisInterval(validPoints.Min(p => p.X), validPoints.Max(p => p.X), out axisMin, out axisMax, out axisInterval);
scatterPlot.VisualProperties.XAxisMinimumAuto = false;
scatterPlot.VisualProperties.XAxisMaximumAuto = false;
scatterPlot.VisualProperties.XAxisMinimumFixedValue = axisMin;
scatterPlot.VisualProperties.XAxisMaximumFixedValue = axisMax;
} catch (ArgumentOutOfRangeException) { } // error during CalculateOptimalAxisInterval
try {
double axisMin, axisMax, axisInterval;
ChartUtil.CalculateOptimalAxisInterval(validPoints.Min(p => p.Y), validPoints.Max(p => p.Y), out axisMin, out axisMax, out axisInterval);
scatterPlot.VisualProperties.YAxisMinimumAuto = false;
scatterPlot.VisualProperties.YAxisMaximumAuto = false;
scatterPlot.VisualProperties.YAxisMinimumFixedValue = axisMin;
scatterPlot.VisualProperties.YAxisMaximumFixedValue = axisMax;
} catch (ArgumentOutOfRangeException) { } // error during CalculateOptimalAxisInterval
}
//No Grouping
if (string.IsNullOrEmpty(variableNameGroup) || variableNameGroup == "-")
{
ScatterPlotDataRow scdr = new ScatterPlotDataRow(variableNameX + " - " + variableNameY, "", validPoints);
scdr.VisualProperties.IsVisibleInLegend = false;
scatterPlot.Rows.Add(scdr);
return(scatterPlot);
}
//Grouping
int groupVariableIndex = preprocessingData.GetColumnIndex(variableNameGroup);
var groupingValues = Enumerable.Empty <string>();
if (preprocessingData.VariableHasType <double>(groupVariableIndex))
{
groupingValues = preprocessingData.GetValues <double>(groupVariableIndex).Select(x => x.ToString());
}
else if (preprocessingData.VariableHasType <string>(groupVariableIndex))
{
groupingValues = preprocessingData.GetValues <string>(groupVariableIndex);
}
else if (preprocessingData.VariableHasType <DateTime>(groupVariableIndex))
{
groupingValues = preprocessingData.GetValues <DateTime>(groupVariableIndex).Select(x => x.ToString());
}
var groups = groupingValues.Zip(validPoints, Tuple.Create).GroupBy(t => t.Item1, t => t.Item2);
if (legendOrder == LegendOrder.Alphabetically)
{
groups = groups.OrderBy(x => x.Key, new NaturalStringComparer());
}
foreach (var group in groups)
{
var scdr = new ScatterPlotDataRow {
Name = group.Key,
VisualProperties =
{
IsVisibleInLegend = true,
PointSize = 6
}
};
scdr.Points.AddRange(group);
scatterPlot.Rows.Add(scdr);
}
return(scatterPlot);
}
private void FillSeriesWithRowValues(Series series, ScatterPlotDataRow row) {
for (int i = 0; i < row.Points.Count; i++) {
var value = row.Points[i];
DataPoint point = new DataPoint();
if (IsInvalidValue(value.X) || IsInvalidValue(value.Y))
point.IsEmpty = true;
else {
point.XValue = value.X;
point.YValues = new double[] { value.Y };
}
series.Points.Add(point);
}
}
private void ConfigureSeries(Series series, ScatterPlotDataRow row) {
series.BorderWidth = 1;
series.BorderDashStyle = ChartDashStyle.Solid;
series.BorderColor = Color.Empty;
if (row.VisualProperties.Color != Color.Empty)
series.Color = row.VisualProperties.Color;
else series.Color = Color.Empty;
series.IsVisibleInLegend = row.VisualProperties.IsVisibleInLegend;
series.ChartType = SeriesChartType.FastPoint;
series.MarkerSize = row.VisualProperties.PointSize;
series.MarkerStyle = ConvertPointStyle(row.VisualProperties.PointStyle);
series.XAxisType = AxisType.Primary;
series.YAxisType = AxisType.Primary;
if (row.VisualProperties.DisplayName.Trim() != String.Empty) series.LegendText = row.VisualProperties.DisplayName;
else series.LegendText = row.Name;
string xAxisTitle = string.IsNullOrEmpty(Content.VisualProperties.XAxisTitle)
? "X"
: Content.VisualProperties.XAxisTitle;
string yAxisTitle = string.IsNullOrEmpty(Content.VisualProperties.YAxisTitle)
? "Y"
: Content.VisualProperties.YAxisTitle;
series.ToolTip =
series.LegendText + Environment.NewLine +
xAxisTitle + " = " + "#VALX," + Environment.NewLine +
yAxisTitle + " = " + "#VAL";
}
public static void PruningChart(RegressionNodeTreeModel tree, ComplexityPruning pruning, ResultCollection results)
{
var nodes = new Queue <RegressionNodeModel>();
nodes.Enqueue(tree.Root);
var max = 0.0;
var strenghts = new SortedList <double, int>();
while (nodes.Count > 0)
{
var n = nodes.Dequeue();
if (n.IsLeaf)
{
max++;
continue;
}
if (!strenghts.ContainsKey(n.PruningStrength))
{
strenghts.Add(n.PruningStrength, 0);
}
strenghts[n.PruningStrength]++;
nodes.Enqueue(n.Left);
nodes.Enqueue(n.Right);
}
if (strenghts.Count == 0)
{
return;
}
var plot = new ScatterPlot("Pruned Sizes", "")
{
VisualProperties =
{
XAxisTitle = "Pruning Strength",
YAxisTitle = "Tree Size",
XAxisMinimumAuto = false,
XAxisMinimumFixedValue = 0
}
};
var row = new ScatterPlotDataRow("TreeSizes", "", new List <Point2D <double> >());
row.Points.Add(new Point2D <double>(pruning.PruningStrength, max));
var fillerDots = new Queue <double>();
var minX = pruning.PruningStrength;
var maxX = strenghts.Last().Key;
var size = (maxX - minX) / 200;
for (var x = minX; x <= maxX; x += size)
{
fillerDots.Enqueue(x);
}
foreach (var strenght in strenghts.Keys)
{
while (fillerDots.Count > 0 && strenght > fillerDots.Peek())
{
row.Points.Add(new Point2D <double>(fillerDots.Dequeue(), max));
}
max -= strenghts[strenght];
row.Points.Add(new Point2D <double>(strenght, max));
}
row.VisualProperties.PointSize = 6;
plot.Rows.Add(row);
results.AddOrUpdateResult("PruningSizes", plot);
}
