rust-automl

automl with smartcore

Train, compare, and predict with machine-learning models in Rust, using SmartCore estimators and configurable preprocessing.

Install

automl = "0.3.1"

Quickstart

Compare classifiers with cross-validation, inspect their scores, and predict with the best model:

use automl::{ClassificationModel, ClassificationSettings, DenseMatrix};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let rows: Vec<Vec<f64>> = (0..24).map(|i| vec![f64::from(i)]).collect();
    let features = DenseMatrix::from_2d_vec(&rows)?;
    let labels: Vec<u32> = (0..24).map(|i| u32::from(i >= 12)).collect();
    let settings = ClassificationSettings::default().with_number_of_folds(3);
    let mut classifier = ClassificationModel::new(features, labels, settings);
    classifier.train()?;

    println!("{classifier}");
    let predictions = classifier.predict(DenseMatrix::from_2d_array(&[&[2.0], &[20.0]])?)?;
    println!("Predictions: {predictions:?}");
    Ok(())
}

A successful train() replaces the previous results. A failed retraining attempt preserves the last successful model and fitted preprocessing; settings edits are retained so you can correct them and retry.