ClusteredBatchStrategy<T, TInput, TOutput>
Models & Types · AiDotNet.ActiveLearning.Batch
Clustering-based batch selection strategy using k-means clustering.
Implements: IClusteringBatchStrategy<T, TInput, TOutput>
For Beginners
This strategy first groups similar samples into clusters, then selects the most informative sample from each cluster. This ensures the batch covers different regions of the data space.
How It Works
How It Works:
Advantages:
Considerations:
Constructors
| Constructor | Summary |
|---|---|
ClusteredBatchStrategy() |
Initializes a new ClusteredBatchStrategy with default settings. |
ClusteredBatchStrategy(int numClusters, int maxIterations, double diversityTradeoff) |
Initializes a new ClusteredBatchStrategy with specified parameters. |
Properties
| Property | Summary |
|---|---|
T DiversityTradeoff { get; set; } |
|
int NumClusters { get; } |
|
string Name { get; } |
Methods
| Method | Summary |
|---|---|
T ComputeDiversity(TInput sample1, TInput sample2) |
|
int[] ClusterSamples(IDataset<T, TInput, TOutput> unlabeledPool) |
|
int[] SelectBatch(int[] candidateIndices, Vector<T> scores, IDataset<T, TInput, TOutput> unlabeledPool, int batchSize) |
|
int[] SelectFromClusters(int[] clusterAssignments, Vector<T> scores, int batchSize) |