Crate dataset_ml 

Built-in dataset implementations for machine learning.

dataset-ml provides ready-to-use loaders for classic ML datasets, built on top of dataset_core::Dataset. Each module is a worked example that shows how to wrap Dataset<T, E> for one data source. The steps are the same each time:

1. Download from a URL.
2. Verify a SHA-256 hash.
3. Parse CSV records, or extract raw documents from an archive.
4. Expose typed accessors backed by ndarray.

Datasets

Module	Samples	Features	Task Type
abalone	4,177	8	Regression
adult	32,561	14	Classification
bank_marketing	45,211	16	Classification
banknote_authentication	1,372	4	Classification
iris	150	4	Classification
breast_cancer	569	30	Classification
boston_housing	506	13	Regression
california_housing	20,640	8	Regression
car_evaluation	1,728	6	Classification
covtype	581,012	54	Classification
diabetes	442	10	Regression
digits	1,797	64	Classification
heart_disease	303	13	Classification
ionosphere	351	34	Classification
kddcup99	494,021 / 4,898,431	41	Classification
letter_recognition	20,000	16	Classification (26 classes)
linnerud	20	3	Regression (multi-output)
mushroom	8,124	22	Classification
spambase	4,601	57	Classification
titanic	891	11	Classification
palmer_penguins	344	7	Classification
sms_spam	5,574	text	Classification
wine_recognition	178	13	Classification
wine_quality::red_wine_quality	1,599	11	Regression
wine_quality::white_wine_quality	4,898	11	Regression
youtube_spam	1,956	text	Classification
sentiment_sentences	3,000	text	Classification
newsgroups20	11,314 / 18,846	text	Classification
movie_review_polarity	2,000	text	Classification

Example

use dataset_ml::iris::Iris;

let iris = Iris::new("./data");
let (features, labels) = iris.data().unwrap();
assert_eq!(features.shape(), &[150, 4]);

All loaders are lazy: the first call downloads and parses the file, every subsequent call returns a cached reference. See the individual module docs for features, target, sample count, and source.

Beyond the loaders

Two modules apply to every dataset here rather than to one of them:

• preprocessing: seeded train/test and k-fold splits (plain or class-stratified), feature scaling, one-hot encoding, and label encoding. You can feed the arrays a loader returns straight to a model, without writing that glue code by hand.
• traits: the MlDataset trait every loader implements. It lets you write code generically over “some dataset”: cache inspection and invalidation, plus a uniform n_samples().

use dataset_ml::preprocessing::{stratified_split, standardize};
use dataset_ml::traits::MlDataset;
use dataset_ml::Iris;
use ndarray::Axis;

let iris = Iris::new("./data");
let (features, labels) = iris.data().unwrap();

// Split with each species proportionally represented on both sides.
let (train, test) = stratified_split(labels.as_slice().unwrap(), 0.2, 42).unwrap();
let (scaled_train, scaler) = standardize(&features.select(Axis(0), &train)).unwrap();

assert_eq!(scaled_train.nrows(), 120);
assert_eq!(iris.n_samples().unwrap(), 150); // from the `MlDataset` trait

Re-exports

pub use abalone::Abalone;

pub use adult::Adult;

pub use bank_marketing::BankMarketing;

pub use banknote_authentication::BanknoteAuthentication;

pub use boston_housing::BostonHousing;

pub use breast_cancer::BreastCancer;

pub use california_housing::CaliforniaHousing;

pub use car_evaluation::CarEvaluation;

pub use covtype::Covtype;

pub use diabetes::Diabetes;

pub use digits::Digits;

pub use heart_disease::HeartDisease;

pub use ionosphere::Ionosphere;

pub use iris::Iris;

pub use kddcup99::Kddcup99;

pub use letter_recognition::LetterRecognition;

pub use linnerud::Linnerud;

pub use movie_review_polarity::MovieReviewPolarity;

pub use mushroom::Mushroom;

pub use newsgroups20::Newsgroups20;

pub use palmer_penguins::PalmerPenguins;

pub use sentiment_sentences::SentimentSentences;

pub use sms_spam::SmsSpam;

pub use spambase::Spambase;

pub use titanic::Titanic;

pub use traits::MlDataset;

pub use traits::NumSamples;

pub use wine_quality::red_wine_quality::RedWineQuality;

pub use wine_quality::white_wine_quality::WhiteWineQuality;

pub use wine_recognition::WineRecognition;

pub use youtube_spam::YoutubeSpam;

Modules

abalone

Abalone dataset module.

adult

Adult / Census Income dataset module.

bank_marketing

Bank Marketing dataset module.

banknote_authentication

Banknote Authentication dataset module.

boston_housing

Boston Housing dataset module.

breast_cancer

Breast Cancer Wisconsin (Diagnostic) dataset module.

california_housing

California Housing dataset module.

car_evaluation

Car Evaluation dataset module.

covtype

Forest Cover Type dataset module.

diabetes

Diabetes dataset module.

digits

Optical Recognition of Handwritten Digits dataset module.

heart_disease

Heart Disease (Cleveland) dataset module.

ionosphere

Ionosphere dataset module.

iris

Iris flower dataset module.

kddcup99

KDD Cup 1999 network-intrusion dataset module.

letter_recognition

Letter Recognition dataset module.

linnerud

Linnerud dataset module.

movie_review_polarity

Movie Review Polarity dataset module.

mushroom

Mushroom dataset module.

newsgroups20

20 Newsgroups dataset module.

palmer_penguins

Palmer Penguins dataset module.

preprocessing

Preprocessing helpers.

sentiment_sentences

Sentiment Labelled Sentences dataset module.

sms_spam

SMS Spam Collection dataset module.

spambase

Spambase dataset module.

titanic

Titanic dataset module.

traits

The traits::MlDataset trait implemented by every loader in this crate.

wine_quality

Wine Quality dataset module.

wine_recognition

Wine Recognition dataset module.

youtube_spam

YouTube Spam Collection dataset module.

Constants

DOWNLOAD_RETRIES

How many extra download attempts every loader in this crate makes before it stops retrying.