COPKMeansOptions<T>
Options & Configuration · AiDotNet.Clustering.Options
Configuration options for COP-KMeans (Constrained K-Means).
Inheritance: ModelOptions → ClusteringOptions<T> → COPKMeansOptions<T>
For Beginners
Sometimes you have partial knowledge about your data.
You might know:
- “These two customers bought the same product” (must-link)
- “These two users are different people” (cannot-link)
COP-KMeans uses this information to guide clustering:
- It won’t separate must-linked pairs
- It won’t group cannot-linked pairs
This is “semi-supervised” because you have some labels but not all.
How It Works
COP-KMeans extends K-Means with pairwise constraints:
- Must-link: Two points must be in the same cluster
- Cannot-link: Two points must be in different clusters
Constructors
| Constructor | Summary |
|---|---|
COPKMeansOptions() |
Initializes COPKMeansOptions with appropriate defaults. |
Properties
| Property | Summary |
|---|---|
List<ValueTuple<int, int>> CannotLink { get; set; } |
Gets or sets the cannot-link constraints. |
List<ValueTuple<int, int>> MustLink { get; set; } |
Gets or sets the must-link constraints. |
bool UseTransitiveClosure { get; set; } |
Gets or sets whether to use transitive closure for must-link constraints. |
int NumClusters { get; set; } |
Gets or sets the number of clusters. |