ferrolearn-linear

Linear models for the ferrolearn machine learning framework.

Algorithms

Regression

Model	Description
`LinearRegression`	Ordinary Least Squares via QR decomposition
`Ridge`	L2-regularized regression via Cholesky decomposition
`Lasso`	L1-regularized regression via coordinate descent
`ElasticNet`	Combined L1/L2 regularization via coordinate descent
`BayesianRidge`	Bayesian ridge with automatic regularization tuning
`HuberRegressor`	Robust regression via IRLS with Huber loss
`SGDRegressor`	Stochastic gradient descent regressor

Classification

Model	Description
`LogisticRegression`	Binary and multiclass classification via L-BFGS
`LDA`	Linear Discriminant Analysis
`SGDClassifier`	Stochastic gradient descent classifier

Example

use ferrolearn_linear::{Ridge, FittedRidge};
use ferrolearn_core::{Fit, Predict};
use ndarray::array;

let x = array![[1.0_f64, 2.0], [3.0, 4.0], [5.0, 6.0]];
let y = array![1.0, 2.0, 3.0];

let model = Ridge::<f64>::new().with_alpha(1.0);
let fitted = model.fit(&x, &y).unwrap();
let predictions = fitted.predict(&x).unwrap();

All models follow the compile-time safety pattern: unfitted structs implement Fit, fitted structs implement Predict. Calling predict() on an unfitted model is a compile error.

Float generics

All models are generic over F: Float + Send + Sync + 'static, supporting both f32 and f64.

License

Licensed under either of Apache License, Version 2.0 or MIT License at your option.