Struct MLPipeline 

pub struct MLPipeline { /* private fields */ }

ML pipeline for data preprocessing and preparation

Implementations

impl MLPipeline

pub fn new(config: MLPipelineConfig) -> Self

Create a new ML pipeline

pub fn prepare_dataset(&mut self, dataset: &Dataset) -> Result<Dataset>

Prepare dataset for ML training

Examples found in repository

examples/real_world_datasets.rs (line 321)

fn demonstrate_advanced_operations() -> Result<(), Box<dyn std::error::Error>> {
    println!("🔧 ADVANCED DATASET OPERATIONS");
    println!("{}", "-".repeat(40));

    let housing = load_california_housing()?;

    // Data preprocessing pipeline
    println!("Preprocessing pipeline for California Housing:");

    // 1. Train/test split
    let (mut train, test) = train_test_split(&housing, 0.2, Some(42))?;
    println!(
        "  1. Split: {} train, {} test",
        train.n_samples(),
        test.n_samples()
    );

    // 2. Feature scaling
    let mut pipeline = MLPipeline::default();
    train = pipeline.prepare_dataset(&train)?;
    println!("  2. Standardized features");

    // 3. Cross-validation setup
    let cv_folds = k_fold_split(train.n_samples(), 5, true, Some(42))?;
    println!("  3. Created {} CV folds", cv_folds.len());

    // Feature correlation analysis (simplified)
    println!("  4. Feature analysis:");
    println!("     • {} numerical features", train.n_features());
    println!("     • Ready for machine learning models");

    // Custom dataset configuration
    println!("\nCustom dataset loading configuration:");
    let config = RealWorldConfig {
        use_cache: true,
        download_if_missing: false, // Don't download in demo
        return_preprocessed: true,
        subset: Some("small".to_string()),
        random_state: Some(42),
        ..Default::default()
    };

    println!("  • Caching: {}", config.use_cache);
    println!("  • Download missing: {}", config.download_if_missing);
    println!("  • Preprocessed: {}", config.return_preprocessed);
    println!("  • Subset: {:?}", config.subset);

    println!();
    Ok(())
}

pub fn train_test_split(&self, dataset: &Dataset) -> Result<DataSplit>

Split dataset into train/test sets

pub fn cross_validation_split( &self, dataset: &Dataset, ) -> Result<CrossValidationFolds>

Generate cross-validation folds

pub fn transform(&self, dataset: &Dataset) -> Result<Dataset>

Transform new data using fitted scalers

pub fn create_experiment(&self, name: &str, dataset: &Dataset) -> MLExperiment

Create an ML experiment tracker

pub fn evaluate_with_cv<F>( &self, dataset: &Dataset, train_fn: F, ) -> Result<CrossValidationResults>

where F: Fn(&Array2<f64>, &Array1<f64>, &Array2<f64>, &Array1<f64>) -> Result<(f64, f64, f64)>,

Evaluate model performance with cross-validation

Trait Implementations

impl Default for MLPipeline

fn default() -> Self

Returns the “default value” for a type.