TabMClassifier<T>

Models & Types ยท AiDotNet.NeuralNetworks.Tabular

TabM implementation for classification tasks.

Inheritance: TabMBase<T> โ†’ TabMClassifier<T>

For Beginners

Use this model when you want to classify data into categories with the benefits of an ensemble (multiple models voting).

How predictions work:

  1. Input passes through shared layers with per-member modulation
  2. Each ensemble member produces class logits
  3. Logits are averaged across members
  4. Softmax converts to probabilities

Benefits over single models:

  • Reduced overfitting through ensemble averaging
  • Better calibrated probability estimates
  • More robust to noisy features
  • Comparable speed to single models (parameter sharing)

Example:

How It Works

TabMClassifier uses the TabM architecture with BatchEnsemble layers for multi-class classification. It averages predictions across ensemble members and uses softmax for probability outputs.

Constructors

Constructor Summary
TabMClassifier() Initializes a new instance of the TabMClassifier class with default configuration.
TabMClassifier(int numFeatures, int numClasses, TabMOptions<T> options) Initializes a new instance of the TabMClassifier class.

Properties

Property Summary
int NumClasses { get; } Gets the number of output classes.

Methods

Method Summary
T ComputeAccuracy(Tensor<T> features, int[] targets) Computes classification accuracy.
T ComputeCrossEntropyLoss(Tensor<T> probabilities, int[] targets) Computes the cross-entropy loss.
T ComputeF1Score(Tensor<T> features, int[] targets) Computes the macro F1 score across all classes.
T TrainStep(Tensor<T> features, int[] targets, T learningRate) Performs a single training step.
Tensor<T> Forward(Tensor<T> features) Performs the forward pass to get class logits (per member).
Tensor<T> PredictProbabilities(Tensor<T> features) Predicts class probabilities (averaged across ensemble members).
Vector<T> GetPredictiveUncertainty(Tensor<T> features) Gets predictive uncertainty as the entropy of the averaged predictions.
int[] Predict(Tensor<T> features) Predicts the most likely class for each sample.
void ResetState() Resets internal state.
void UpdateParameters(T learningRate) Updates all parameters including the classification head.