SVM.NuSVC

Variant of SVC that uses the nu parameter to control the fraction of support vectors and the margin errors.

interface NuSVCProps {
    nu?: number;
}
constructor(props: NuSVCProps = {})

Parameters

• nu (number, default 0.5): trade-off between the number of support vectors and training errors
• C (number, default 1/nu): regularization strength
• maxIter (number, default 100): maximum iterations for training
• learningRate (number, default 0.01): step size for gradient descent
• kernel ('linear' | 'rbf', default 'rbf'): kernel type
• gamma (number, default 1): RBF kernel coefficient

NuSVC shares the same usage as SVC but uses the nu parameter instead of C.

Practical guide: NuSVC in JavaScript and TypeScript

NuSVC provides SVM classification with nu-based control over support vector fraction and margin errors.

When to use NuSVC

• You prefer nu constraints over classic C tuning behavior.
• Dataset boundaries require flexible kernel-based classification.
• You need explicit control over support-vector complexity.

Implementation workflow

1. Choose kernel and initialize a conservative nu value.
2. Fit NuSVC and inspect validation accuracy and support vector count.
3. Tune kernel and nu jointly for generalization and runtime limits.

JavaScript deployment notes

• Prefer feature scaling for distance-based and gradient-based algorithms to improve stability.
• In browser apps, run heavy training in Web Workers to keep UI interactions smooth.
• Keep a simple baseline from the same module as a fallback model for comparison.

Search intents this page targets

• NuSVC JavaScript
• NuSVC TypeScript
• NuSVC browser machine learning
• @kanaries/ml NuSVC