MissingFeatureOptions
Options & Configuration · AiDotNet.Models.Options
Configuration for handling missing features in vertical federated learning.
For Beginners
In real-world VFL deployments, not all parties have data for all entities. For example, a bank might have records for 100,000 customers, but the partnering hospital only has records for 30,000 of those patients. The other 70,000 customers have “missing” hospital features.
How It Works
This class controls how those missing features are handled during training and inference.
Example:
Constructors
| Constructor | Summary |
|---|---|
MissingFeatureOptions() |
Properties
| Property | Summary |
|---|---|
MissingFeatureStrategy Strategy { get; set; } |
Gets or sets the strategy for imputing missing features. |
bool AddMissingnessIndicator { get; set; } |
Gets or sets whether to create a binary indicator feature marking which features are imputed vs. |
bool AllowPartialAlignment { get; set; } |
Gets or sets whether to include partially-aligned entities in training. |
double AlignmentThreshold { get; set; } |
Gets or sets the alignment confidence threshold when using fuzzy entity matching. |
double MinimumOverlapRatio { get; set; } |
Gets or sets the minimum required overlap ratio between parties before training can proceed. |