// O(k*n)
List <XY> RunClusterAlgo()
{
// put points in buckets
int allPointsCount = BaseDataset.Count;
var firstPoint = BaseDataset[0];
firstPoint.Color = GetRandomColor();
var firstId = 0.ToString();
var firstBucket = new Bucket(firstId)
{
Centroid = firstPoint
};
BaseBucketsLookup.Add(firstId, firstBucket);
for (int i = 1; i < allPointsCount; i++)
{
var set = new HashSet <string>(); //cluster candidate list
var p = BaseDataset[i];
// iterate clusters and collect candidates
foreach (var bucket in BaseBucketsLookup.Values)
{
var isInCluster = MathTool.DistWithin(p, bucket.Centroid, DISTANCECLUSTERER_SIZE);
if (!isInCluster)
{
continue;
}
set.Add(bucket.Id);
//use first, short dist will be calc at last step before returning data
break;
}
// if not within box area, then make new cluster
if (set.Count == 0)
{
var pid = i.ToString();
p.Color = GetRandomColor();
var newbucket = new Bucket(pid)
{
Centroid = p
};
BaseBucketsLookup.Add(pid, newbucket);
}
}
//important, align all points to closest cluster point
BaseUpdatePointsByCentroid();
return(BaseGetClusterResult());
}
