NODEBase<T>
Base Classes · AiDotNet.NeuralNetworks.Tabular
Base class for NODE (Neural Oblivious Decision Ensembles).
Implements: IParameterSource<T>
Known derived types: NODEClassifier<T>, NODERegression<T>
For Beginners
Think of NODE as making decision trees trainable like neural networks:
- Oblivious trees: At each depth, all nodes split on the same feature.
This makes trees faster and more regularized.
- Soft splits: Instead of “if feature > threshold then go right”,
we use a smooth function that gradually transitions.
- Ensemble: Multiple trees vote together for the final answer.
The result is a model that combines tree interpretability with deep learning power.
How It Works
NODE uses differentiable oblivious decision trees that can be trained end-to-end:
- Each tree uses the same feature at each depth level (oblivious)
- Split decisions are soft (differentiable) using sigmoid
- Trees are combined additively for the final output
Properties
| Property | Summary |
|---|---|
int TreeOutputDimension { get; } |
Gets the tree output dimension. |
long ParameterCount { get; } |
Methods
| Method | Summary |
|---|---|
Vector<T> GetFeatureImportance() |
Gets the feature importance scores. |
Vector<T> GetParameters() |
Reads every parameter in traversal order. |
void ResetState() |
Resets internal state. |
void SetParameters(Vector<T> parameters) |
Restores every parameter, in the order GetParameters emitted them. |
void UpdateParameters(T learningRate) |
Updates all parameters. |