DARTClassifierOptions<T>

Options & Configuration · AiDotNet.Models.Options

Configuration options for DART (Dropouts meet Multiple Additive Regression Trees) classifier.

Inheritance: ModelOptions → ClassifierOptions<T> → DARTClassifierOptions<T>

For Beginners

DART is gradient boosting with dropout - it randomly “forgets” some trees during training to prevent overfitting. This is similar to dropout in neural networks.

Key options:

  • DropoutRate: Fraction of trees to drop each iteration (higher = more regularization)
  • DropoutType: How to select which trees to drop
  • NormalizationType: How to scale predictions after dropout

How It Works

DART applies dropout regularization to gradient boosting classification. During each iteration, a random subset of existing trees is dropped, helping prevent overfitting.

Constructors

Constructor Summary
DARTClassifierOptions()

Properties

Property Summary
DARTDropoutType DropoutType { get; set; } Gets or sets the type of dropout selection.
DARTNormalizationType NormalizationType { get; set; } Gets or sets the normalization type after dropout.
SplitCriterion SplitCriterion { get; set; } Gets or sets the split criterion for the base trees.
bool Verbose { get; set; } Gets or sets whether to print verbose output.
double DropoutRate { get; set; } Gets or sets the dropout rate - fraction of trees to drop each iteration.
double LearningRate { get; set; } Gets or sets the learning rate (shrinkage factor).
double MaxFeatures { get; set; } Gets or sets the fraction of features to consider for splits.
int MaxDepth { get; set; } Gets or sets the maximum depth of trees.
int MinSamplesLeaf { get; set; } Gets or sets the minimum number of samples at a leaf node.
int MinSamplesSplit { get; set; } Gets or sets the minimum number of samples required to split.
int NumberOfIterations { get; set; } Gets or sets the number of boosting iterations.
int VerboseEval { get; set; } Gets or sets how often to print progress.
int? EarlyStoppingRounds { get; set; } Gets or sets early stopping rounds.