LinearSVC
Support Vector Machine (SVM) Classifier.
Implements a linear SVM using sub-gradient descent on the hinge loss with L2 regularization (soft margin). Suitable for binary classification tasks.
Mathematical Formulation:
- Decision function: f(x) = sign(w · x + b)
- Optimization: minimize (1/2)||w||² + C * Σmax(0, 1 - y_i(w · x_i + b))
Constructor
new LinearSVC(options?: {
C?: number;
maxIter?: number;
tol?: number;
})
Regularization parameter (inverse of regularization strength). Larger C = stronger penalty on errors = harder margin.
Maximum iterations for optimization.
Methods
fit
fit(X: Tensor, y: Tensor): this
Training data of shape (n_samples, n_features)
Target labels of shape (n_samples,). Must contain at least 2 classes.
predict
predict(X: Tensor): Tensor
predictProba
predictProba(X: Tensor): Tensor
score
score(X: Tensor, y: Tensor): number
Properties
Weight vectors as tensor of shape (n_classes, n_features)
Example
import { LinearSVC } from 'deepbox/ml';
import { tensor } from 'deepbox/ndarray';
const X = tensor([[1, 2], [2, 3], [3, 1], [4, 2]]);
const y = tensor([0, 0, 1, 1]);
const svm = new LinearSVC({ C: 1.0 });
svm.fit(X, y);
const predictions = svm.predict(X);
LinearSVR
Support Vector Machine (SVM) Regressor.
Implements epsilon-SVR (Support Vector Regression) using sub-gradient descent.
Constructor
new LinearSVR(options?: {
C?: number;
epsilon?: number;
maxIter?: number;
tol?: number;
})
Regularization parameter. Must be positive.
Epsilon in the epsilon-SVR model. Defines the epsilon-tube within which no penalty is associated in the training loss function.
Maximum number of iterations.
Tolerance for stopping criterion.
Methods
fit
fit(X: Tensor, y: Tensor): this
Training data of shape (n_samples, n_features)
Target values of shape (n_samples,)
predict
predict(X: Tensor): Tensor
score
score(X: Tensor, y: Tensor): number
Example
import { LinearSVR } from 'deepbox/ml';
import { tensor } from 'deepbox/ndarray';
const X = tensor([[1], [2], [3], [4]]);
const y = tensor([1.5, 2.5, 3.5, 4.5]);
const svr = new LinearSVR({ C: 1.0, epsilon: 0.1 });
svr.fit(X, y);
const predictions = svr.predict(X);