SOMOptions<T>
Options & Configuration · AiDotNet.Clustering.Options
Configuration options for Self-Organizing Maps (SOM).
Inheritance: ModelOptions → ClusteringOptions<T> → SOMOptions<T>
For Beginners
SOM creates a “map” of your data.
Imagine compressing a 3D world onto a 2D map:
- Nearby countries on the map should be nearby in reality
- The map preserves relationships even though it’s lower dimensional
SOM does this for high-dimensional data:
- Creates a 2D grid of “neurons”
- Each neuron represents a prototype pattern
- Similar data points activate nearby neurons
Uses:
- Visualization of high-dimensional data
- Dimensionality reduction that preserves topology
- Finding natural groupings in data
How It Works
Self-Organizing Maps are a type of neural network that produces a low-dimensional (typically 2D) discretized representation of the input space. They preserve topological properties of the input, meaning similar inputs map to nearby neurons.
Constructors
| Constructor | Summary |
|---|---|
SOMOptions() |
Initializes SOMOptions with appropriate defaults. |
Properties
| Property | Summary |
|---|---|
NeighborhoodFunction NeighborhoodType { get; set; } |
Gets or sets the neighborhood function type. |
SOMTopology Topology { get; set; } |
Gets or sets the topology of the SOM grid. |
double InitialLearningRate { get; set; } |
Gets or sets the initial learning rate. |
double InitialNeighborhoodRadius { get; set; } |
Gets or sets the initial neighborhood radius. |
int GridHeight { get; set; } |
Gets or sets the height of the SOM grid. |
int GridWidth { get; set; } |
Gets or sets the width of the SOM grid. |