BallTree<T>
Models & Types · AiDotNet.Clustering.SpatialIndex
Ball Tree for efficient nearest neighbor queries with arbitrary distance metrics.
For Beginners
A Ball Tree groups nearby points into “balls” (spheres). Each ball contains a center point and a radius that encloses all points in that ball.
When searching for neighbors:
- If the query point is far from a ball’s center (farther than the ball’s radius
plus your search radius), you can skip the entire ball.
- This saves a lot of computation for large datasets.
Ball Trees work better than KD-Trees when:
- Your data has many dimensions (>20)
- You’re using non-Euclidean distance metrics (like cosine distance)
How It Works
A Ball Tree partitions space using hyperspheres (balls) rather than hyperplanes. This makes it more effective than KD-Trees for high-dimensional data and for non-Euclidean distance metrics.
Constructors
| Constructor | Summary |
|---|---|
BallTree(IDistanceMetric<T> distanceMetric, int leafSize) |
Initializes a new Ball Tree instance. |
Properties
| Property | Summary |
|---|---|
int Count { get; } |
Gets the number of points in the tree. |
int Dimensions { get; } |
Gets the number of dimensions. |
Methods
| Method | Summary |
|---|---|
ValueTuple<int, T>[] QueryKNearest(Vector<T> query, int k) |
Finds the k nearest neighbors to the query point. |
ValueTuple<int, T>[] QueryRadius(Vector<T> query, T radius) |
Finds all points within the given radius of the query point. |
void Build(Matrix<T> data) |
Builds the Ball Tree from the given data. |