public void DetectAnomaliesAndSetResults(string pathCSVAnomalies)
{
if (Results != IntPtr.Zero)
{
LibraryManager.deleteResults(Results);
Results = IntPtr.Zero;
}
IntPtr train = StringWrapper.CreateStringWrapperFromString(mainModel.LearnFile), test = StringWrapper.CreateStringWrapperFromString(pathCSVAnomalies);
Results = LibraryManager.detectAnomalis(train, test);
SetAnomalies(Results);
StringWrapper.removeStr(train);
StringWrapper.removeStr(test);
}
private string GetCorrelatedFeature(string feature)
{
if (Results == IntPtr.Zero)
{
return(null);
}
IntPtr f = StringWrapper.CreateStringWrapperFromString(feature);
IntPtr corr = LibraryManager.getCorrelateFeatureByFeatureName(Results, f);
string corrStr = StringWrapper.GetStr(corr);
StringWrapper.removeStr(corr);
StringWrapper.removeStr(f);
return(corrStr);
}
public void UpdateAxes()
{
string category = mainModel.GetCurrentCategory();
if (CurrCategoryPM.Axes.Count > 1)
{
var yAxis = CurrCategoryPM.Axes.ElementAt(1);
yAxis.Minimum = mainModel.GetCategoryMinimum(category) - 5;
yAxis.Maximum = mainModel.GetCategoryMaximum(category) + 5;
yAxis.Title = category;
}
if (CurrCorrelatedCategoryPM.Axes.Count > 1)
{
var yAxis = CurrCorrelatedCategoryPM.Axes.ElementAt(1);
string corrFeat = GetCorrelatedFeature(category);
yAxis.Minimum = mainModel.GetCategoryMinimum(corrFeat) - 5;
yAxis.Maximum = mainModel.GetCategoryMaximum(corrFeat) + 5;
yAxis.Title = corrFeat;
}
if (CorrelatedAsFuncOfCurrent.Axes.Count > 1)
{
var xAxis = CorrelatedAsFuncOfCurrent.Axes.ElementAt(0);
xAxis.Minimum = mainModel.GetCategoryMinimum(category) - 5;
xAxis.Maximum = mainModel.GetCategoryMaximum(category) + 5;
xAxis.Title = category;
var yAxis = CorrelatedAsFuncOfCurrent.Axes.ElementAt(1);
string corrFeat = GetCorrelatedFeature(category);
yAxis.Minimum = mainModel.GetCategoryMinimum(corrFeat) - 5;
yAxis.Maximum = mainModel.GetCategoryMaximum(corrFeat) + 5;
yAxis.Title = corrFeat;
float a, b;
var feature = StringWrapper.CreateStringWrapperFromString(category);
a = LibraryManager.getSlope(Results, feature);
b = LibraryManager.getYIntercept(Results, feature);
StringWrapper.removeStr(feature);
regLine = new FunctionSeries((double x) => a * x + b, xAxis.Minimum, xAxis.Maximum, 0.1);
// Line reg = LinearRegressionCalculator.CalculateLineRegression(learnedData[category], learnedData[GetCorrelatedFeature(category)]);
// regLine = new FunctionSeries((double x) => reg.Slope * x + reg.YIntercept, xAxis.Minimum, xAxis.Maximum, 0.0001);
}
AnomaliesData?.Clear();
foreach (Anomaly a in anomalies)
{
if (a.GetFeature1() == category)
{
AnomaliesData.Add(a);
}
}
}
private void SetAnomalies(IntPtr results)
{
AnomaliesData?.Clear();
anomalies = new List <Anomaly>();
int len = LibraryManager.anomaliesLen(results);
for (int i = 0; i < len; ++i)
{
IntPtr desc = LibraryManager.getAnomalyDescription(results, i);
int timeStep = LibraryManager.getAnomalyTimeStep(results, i);
anomalies.Add(new Anomaly(timeStep, StringWrapper.GetStr(desc)));
// VM_AnomaliesData.Add(new Anomaly(timeStep, GetStr(desc)));
StringWrapper.removeStr(desc);
}
}
