IFeatureMapper<T>

Interfaces · AiDotNet.TransferLearning.FeatureMapping

Defines the interface for mapping features from a source domain to a target domain.

Known implementations: LinearFeatureMapper<T>

For Beginners

A feature mapper is like a translator between two different languages. When you have data from one domain (source) and want to use it in another domain (target), the feature mapper transforms the data so it makes sense in the new context.

How It Works

For example, if you trained a model on images (which might have thousands of features) and want to use that knowledge for text (which has different features), a feature mapper helps bridge that gap by finding a common representation.

Properties

Property Summary
bool IsTrained { get; } Determines if the mapper has been trained and is ready to use.

Methods

Method Summary
Matrix<T> MapToSource(Matrix<T> targetFeatures, int sourceDimension) Maps features from the target domain back to the source domain.
Matrix<T> MapToTarget(Matrix<T> sourceFeatures, int targetDimension) Maps features from the source domain to the target domain.
T GetMappingConfidence() Gets the confidence score for the mapping quality.
void Train(Matrix<T> sourceData, Matrix<T> targetData) Trains the feature mapper on source and target data.