Table of Contents

Class MonteCarloValidationOptions

Namespace
AiDotNet.Models.Options
Assembly
AiDotNet.dll

Represents the options for Monte Carlo cross-validation.

public class MonteCarloValidationOptions : CrossValidationOptions
Inheritance
MonteCarloValidationOptions
Inherited Members

Remarks

This class extends the base CrossValidationOptions with additional properties specific to Monte Carlo cross-validation.

For Beginners: Monte Carlo cross-validation options help you customize how the Monte Carlo method splits and tests your data.

What this class does:

  • Inherits all the basic cross-validation options (like number of folds)
  • Adds a new option to set the size of the validation set

This is useful because:

  • It allows you to control how much of your data is used for validation in each Monte Carlo iteration
  • You can adjust this to find the right balance between your training and validation set sizes

Think of it like deciding how to split a deck of cards for a card game - this option lets you choose how many cards go into each pile (training and validation) for each round of Monte Carlo testing.

Properties

ValidationSize

Gets or sets the size of the validation set as a proportion of the total dataset.

public double ValidationSize { get; set; }

Property Value

double

A double value between 0 and 1, representing the fraction of the dataset to use for validation. The default value is 0.2 (20% of the data used for validation).

Remarks

This property determines what portion of the data will be used for validation in each Monte Carlo iteration.

For Beginners: This sets how much of your data is used to test your model in each round.

What it does:

  • A value of 0.2 means 20% of your data will be used for testing (validation) and 80% for training
  • You can adjust this value to change the split between training and validation data

For example:

  • If you have 1000 data points and set this to 0.2, each round will use 200 points for testing and 800 for training
  • If you set it to 0.3, each round will use 300 points for testing and 700 for training

Choose a value that gives you enough data for both training and testing your model effectively.

Edit this page