GMeans<T>
Models & Types · AiDotNet.Clustering.AutoK
G-Means clustering with automatic K determination using Gaussian tests.
Inheritance: ClusteringBase<T> → GMeans<T>
For Beginners
G-Means assumes good clusters are Gaussian.
A Gaussian distribution is the famous “bell curve”:
- Most points near the center
- Fewer points as you go farther out
- Symmetric left and right
If a cluster doesn’t look like a bell curve:
- It might contain multiple groups
- Split it into two and check again
This works well when true clusters are roughly Gaussian.
How It Works
G-Means tests whether data in each cluster comes from a Gaussian distribution. If not, the cluster is split. It uses the Anderson-Darling test projected onto the principal direction connecting cluster centers.
Algorithm:
- Start with K=1 (all data in one cluster)
- For each cluster, project data onto the main axis
- Test if projection is Gaussian (Anderson-Darling)
- If not Gaussian, split the cluster
- Repeat until all clusters are Gaussian or max K reached
Example
using AiDotNet;
using AiDotNet.Data.Loaders;
using AiDotNet.Clustering.AutoK;
using AiDotNet.Tensors.LinearAlgebra;
var data = new Matrix<double>(6, 2);
double[][] rows = { new[] { 1.0, 1.0 }, new[] { 1.2, 0.9 }, new[] { 1.1, 1.1 },
new[] { 8.0, 8.0 }, new[] { 8.2, 7.9 }, new[] { 7.9, 8.1 } };
for (int i = 0; i < 6; i++) { data[i, 0] = rows[i][0]; data[i, 1] = rows[i][1]; }
var result = await new AiModelBuilder<double, Matrix<double>, Vector<double>>()
.ConfigureModel(new GMeans<double>())
.ConfigureDataLoader(DataLoaders.FromMatrix(data))
.BuildAsync();
var labels = result.Predict(data);
Console.WriteLine($"GMeans: clustered {labels.Length} points.");
Constructors
| Constructor | Summary |
|---|---|
GMeans(GMeansOptions<T> options) |
Initializes a new GMeans instance. |
Methods
| Method | Summary |
|---|---|
IFullModel<T, Matrix<T>, Vector<T>> WithParameters(Vector<T> parameters) |
|
ModelOptions GetOptions() |
|
Vector<T> FitPredict(Matrix<T> x) |
|
Vector<T> Predict(Matrix<T> x) |
|
void Train(Matrix<T> x, Vector<T> y) |