datarust 0.6.5

Scikit-learn-style preprocessing and classical ML in Rust
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34

//! Regression estimators with a `fit`/`predict` API.
//!
//! Provides ordinary least squares ([`LinearRegression`]), L2-regularized
//! ([`Ridge`]), L1-regularized ([`Lasso`]) linear models, and binary
//! [`LogisticRegression`] for classification. All share the
//! [`Predictor`](crate::traits::Predictor) trait and the [`crate::linalg`]
//! solver foundation; regression models additionally implement
//! [`Regressor`](crate::traits::Regressor), while logistic regression implements
//! [`Classifier`](crate::traits::Classifier).

/// L1-regularized regression (coordinate descent).
pub mod lasso;
/// Ordinary least-squares linear regression.
pub mod linear_regression;
/// Binary logistic regression (IRLS solver).
pub mod logistic_regression;
/// L2-regularized ridge regression.
pub mod ridge;

pub use lasso::Lasso;
pub use linear_regression::{LinearRegression, LinearSolver};
pub use logistic_regression::{LogisticRegression, LogisticSolver};
pub use ridge::{Ridge, RidgeSolver};

pub(crate) fn validate_finite_targets(y: &[f64]) -> crate::error::Result<()> {
    for (index, &value) in y.iter().enumerate() {
        if !value.is_finite() {
            return Err(crate::error::DatarustError::InvalidInput(format!(
                "target at index {index} must be finite, found {value}"
            )));
        }
    }
    Ok(())
}