public void LearnHelper(TimeSeries ts, float p /*pearson*/, string f1, string f2, Point[] ps)
{
if (p > threshold)
{
int len = ts.getNumOfTimesteps();
correlatedFeatures c = new correlatedFeatures(f1, f2, p, AnomalyDetectionUtil.LinearReg(ps, len), findThreshold(ps, len, AnomalyDetectionUtil.LinearReg(ps, len)) * (float)1.1);
cf.Add(c);
}
}
public void UpdateAnomaly(List <AnomalyReport> ar, Point p, correlatedFeatures cor, int timeStepIndex, string f1, string f2)
{
if (AnomalyDetectionUtil.Dev(p, cor.LineReg) > cor.Threshhold)
{
string desc = f1 + "-" + f2;
long timeS = timeStepIndex + 1;
AnomalyReport rep = new AnomalyReport(desc, timeS);
ar.Add(rep);
}
}
public void LearnHelper(TimeSeries ts, float p /*pearson*/, string f1, string f2, Point[] ps)
{
base.LearnHelper(ts, p, f1, f2, ps);
if (p > 0.5 && p < maxC)
{
MinCircle circ = new MinCircle();
Circle circle = circ.findMinCircle(ps, ts.getNumOfTimesteps());
correlatedFeatures c = new correlatedFeatures(f1, f2, p, null, circle.radius * (float)1.1);
c.CX = circle.center.x;
c.CY = circle.center.y;
base.CF.Add(c);
}
}
public void UpdateAnomaly(List <AnomalyReport> reportList, Point p, correlatedFeatures corF, int timeStepIndex, string f1, string f2)
{
// in case the correlated features are fulfill the condition of the parent class
if (corF.Correlation >= maxC)
{
base.UpdateAnomaly(reportList, p, corF, timeStepIndex, f1, f2);
}
else
{
Circle minCircle = new Circle(new Point(corF.CX, corF.CY), corF.Threshhold);
MinCircle minHelperCirc = new MinCircle();
if (!minHelperCirc.isPointInsideCircle(p, minCircle))
{
// we will report as anomaly
string descr = f1 + "-" + f2;
long timeS = timeStepIndex + 1;
AnomalyReport rep = new AnomalyReport(descr, timeS);
reportList.Add(rep);
}
}
}
