Table of Contents

Interface ICrossValidator<T, TInput, TOutput>

Namespace
AiDotNet.Interfaces
Assembly
AiDotNet.dll

Defines the contract for cross-validation implementations in machine learning models.

public interface ICrossValidator<T, TInput, TOutput>

Type Parameters

T

The numeric type used for calculations (e.g., float, double, decimal).

TInput

The type of input data (e.g., Matrix<T> for tabular data, custom types for other formats).

TOutput

The type of output data (e.g., Vector<T> for predictions, custom types for other formats).

Remarks

This interface specifies the method signature for performing cross-validation on machine learning models. Cross-validation is a crucial technique for assessing how the results of a statistical analysis will generalize to an independent data set. It's particularly important in contexts where the goal is prediction, and one wants to estimate how accurately a predictive model will perform in practice.

For Beginners: This interface is like a blueprint for creating cross-validation tools.

What it does:

  • Defines a standard way to perform cross-validation on any machine learning model
  • Ensures that all cross-validation implementations will work the same way, regardless of the specific details
  • Works with any data format (matrices, tensors, custom structures) through generic type parameters

It's like setting a standard recipe that all cross-validation methods must follow, ensuring consistency and ease of use across different types of models and data.

Methods

Validate(IFullModel<T, TInput, TOutput>, TInput, TOutput, IOptimizer<T, TInput, TOutput>)

Performs cross-validation on the given model using the provided data and optimizer.

CrossValidationResult<T, TInput, TOutput> Validate(IFullModel<T, TInput, TOutput> model, TInput X, TOutput y, IOptimizer<T, TInput, TOutput> optimizer)

Parameters

model IFullModel<T, TInput, TOutput>

The machine learning model to validate.

X TInput

The input data containing the features.

y TOutput

The output data containing the targets.

optimizer IOptimizer<T, TInput, TOutput>

The optimizer to use for training the model on each fold.

Returns

CrossValidationResult<T, TInput, TOutput>

A CrossValidationResult containing the results of the validation process.

Remarks

This method implements the core cross-validation logic. It typically involves splitting the data into folds, training and evaluating the model on different combinations of these folds using the provided optimizer, and aggregating the results. The specific implementation details may vary depending on the type of cross-validation being performed.

For Beginners: This method is where the actual cross-validation happens.

What it does:

  • Takes your model, your data (X and y), and an optimizer for training
  • Splits your data into parts based on your options
  • Trains your model using the optimizer multiple times on different parts
  • Tests the trained model on held-out data for each fold
  • Collects and summarizes the results of all these tests

The optimizer parameter allows you to use advanced optimization techniques (like genetic algorithms, Bayesian optimization, etc.) during cross-validation, ensuring consistent training across all folds.

It's like putting your model through a series of tests using a standardized training procedure and then giving you a report card that shows how well it performed overall.

Edit this page