ClientSelectionOptions
Options & Configuration ยท AiDotNet.Models.Options
Configuration options for client selection in federated learning.
Inheritance: ModelOptions โ ClientSelectionOptions
How It Works
For Beginners: Client selection controls which devices/organizations participate in each round. In real deployments, many clients may be offline or slow, so selecting a subset per round is common.
Constructors
| Constructor | Summary |
|---|---|
ClientSelectionOptions() |
Properties
| Property | Summary |
|---|---|
Dictionary<int, double> ClientAvailabilityProbabilities { get; set; } |
Gets or sets an optional mapping from client ID to an availability probability (0.0 to 1.0). |
Dictionary<int, string> ClientGroupKeys { get; set; } |
Gets or sets an optional mapping from client ID to a group key for stratified sampling. |
FederatedClientSelectionStrategy Strategy { get; set; } |
Gets or sets the selection strategy. |
double AvailabilityThreshold { get; set; } |
Gets or sets the minimum availability probability required for a client to be considered available. |
double ExplorationRate { get; set; } |
Gets or sets the exploration probability for performance-aware sampling (0.0 to 1.0). |
int ClusterCount { get; set; } |
Gets or sets the number of clusters to use for cluster-based sampling. |
int KMeansIterations { get; set; } |
Gets or sets the number of k-means iterations for cluster-based sampling. |