Struct Dataset 

pub struct Dataset<X = ArrayBase<OwnedRepr<f64>, Dim<[usize; 2]>>, Y = ArrayBase<OwnedRepr<f64>, Dim<[usize; 1]>>> {
    pub data: X,
    pub target: Y,
    pub feature_names: Vec<String>,
    pub target_names: Option<Vec<String>>,
    pub description: String,
}

A simple dataset structure for machine learning operations

The Dataset struct is the primary data container for sklears, holding feature matrices and target values along with metadata.

Type Parameters

• X: Type of the feature matrix (defaults to Array2``)
• Y: Type of the target values (defaults to Array1``)

Examples

use sklears_core::dataset::Dataset;
use scirs2_core::ndarray::{Array1, Array2};

let features = Array2::`<f64>`::zeros((100, 4));
let targets = Array1::`<f64>`::zeros(100);
let dataset = Dataset::new(features, targets)
    .with_description("Sample dataset".to_string());

Fields

data: X

Feature matrix (n_samples x n_features)

target: Y

Target values (n_samples,)

feature_names: Vec<String>

Feature names for interpretability

target_names: Option<Vec<String>>

Target names (for classification tasks)

description: String

Dataset description for documentation

Implementations

impl<X, Y> Dataset<X, Y>

pub fn new(data: X, target: Y) -> Dataset<X, Y>

Create a new dataset with the given data and target

This is the primary constructor for creating datasets. Additional metadata can be added using builder methods.

Arguments

• data - Feature matrix or data structure
• target - Target values corresponding to the features

Returns

A new Dataset instance with empty metadata

pub fn builder() -> DatasetBuilder<X, Y, NoData, NoTarget>

Create a builder for constructing a dataset with compile-time validation

The builder pattern provides compile-time guarantees that both data and targets are provided before the dataset can be constructed.

Returns

A DatasetBuilder in its initial state

pub fn with_feature_names(self, names: Vec<String>) -> Dataset<X, Y>

Set feature names for the dataset

Feature names improve interpretability and are used in various visualization and analysis tools.

Arguments

• names - Vector of feature names

Returns

Self with updated feature names

pub fn with_target_names(self, names: Vec<String>) -> Dataset<X, Y>

Set target names for classification tasks

Target names are particularly useful for multi-class classification where class labels need to be interpretable.

Arguments

• names - Vector of class/target names

Returns

Self with updated target names

pub fn with_description(self, description: String) -> Dataset<X, Y>

Set a description for the dataset

Descriptions are useful for documenting the source, preprocessing steps, or other relevant information about the dataset.

Arguments

• description - String description of the dataset

Returns

Self with updated description

pub fn n_samples(&self) -> Option<usize>

where X: HasShape,

Get the number of samples in the dataset

This is a convenience method that should be implemented by types that can determine their sample count.

pub fn n_features(&self) -> Option<usize>

where X: HasShape,

Get the number of features in the dataset

This is a convenience method that should be implemented by types that can determine their feature count.

Trait Implementations

impl<X, Y> Clone for Dataset<X, Y>

where X: Clone, Y: Clone,

fn clone(&self) -> Dataset<X, Y>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<X, Y> Debug for Dataset<X, Y>

where X: Debug, Y: Debug,

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl StableApi for Dataset

const STABLE_SINCE: &'static str = "0.1.0"

API version this type was stabilized in

const HAS_EXPERIMENTAL_FEATURES: bool = false

Whether this API has any experimental features