/// <summary>
/// Radius search.
/// </summary>
///
private void nearest(TNode current, double[] position,
double radius, ICollection <NodeDistance <TNode> > list)
{
// Check if the distance of the point from this
// node is within the desired radius, and if it
// is, add to the list of nearest nodes.
double d = distance.Distance(position, current.Position);
if (d <= radius)
{
list.Add(new NodeDistance <TNode>(current, d));
}
// Prepare for recursion. The following null checks
// will be used to avoid function calls if possible
double value = position[current.Axis];
double median = current.Position[current.Axis];
double u = value - median;
if (u <= 0)
{
if (current.Left != null)
{
nearest(current.Left, position, radius, list);
}
if (current.Right != null && Math.Abs(u) <= radius)
{
nearest(current.Right, position, radius, list);
}
}
else
{
if (current.Right != null)
{
nearest(current.Right, position, radius, list);
}
if (current.Left != null && Math.Abs(u) <= radius)
{
nearest(current.Left, position, radius, list);
}
}
}
public void DistanceUnitTest()
{
Assert.Inconclusive("TODO");
IMetric target = CreateIMetric(); // TODO: Initialize to an appropriate value
IMBP myPoint1 = null; // TODO: Initialize to an appropriate value
IMBP myPoint2 = null; // TODO: Initialize to an appropriate value
float expected = 0F; // TODO: Initialize to an appropriate value
float actual;
actual = target.Distance(myPoint1, myPoint2);
Assert.AreEqual(expected, actual);
}
private void ComputeCentroid(IEnumerable <IEntitySpecification> items, IMetricItem mean)
{
var minDist = double.MaxValue;
foreach (var item in items)
{
var currentDistance = _metric.Distance(item, mean);
if (!(currentDistance < minDist))
{
continue;
}
minDist = currentDistance;
Height = item.Height;
Weight = item.Weight;
}
}
public static string Classify(ReutersMetricObject reuter, IEnumerable <ReutersMetricObject> trainingData, IMetric metric)
{
var distance = new SortedList <double, string>(new DuplicateComparer <double>());
var categories = new Dictionary <string, int>();
foreach (var o in trainingData)
{
distance.Add(metric.Distance(reuter.Vector, o.Vector), o.Property);
}
for (int i = 0; i < Settings.Default.K; i++)
{
categories.TryGetValue(distance.ElementAt(i).Value, out var value);
categories[distance.ElementAt(i).Value] = value + 1;
}
return(categories.First(c => c.Value == categories.Max(x => x.Value)).Key);
}
private bool Assign()
{
var numClusters = _clusters.Count;
var isChanged = false;
var distances = new double[numClusters];
foreach (var t in _items)
{
for (var k = 0; k < numClusters; k++)
{
distances[k] = _metric.Distance(t, _clusters[k].Centroid);
}
var newClusterId = _metric.MinIndex(distances);
if (newClusterId == t.ClusterId)
{
continue;
}
isChanged = true;
t.SetCluster(newClusterId);
}
return(isChanged);
}