sklears-preprocessing

Latest release: 0.1.0-beta.1 (January 1, 2026). See the workspace release notes for highlights and upgrade guidance.

Overview

sklears-preprocessing contains scalers, encoders, transformers, and feature engineering utilities that mirror scikit-learn’s preprocessing module while leveraging Rust performance.

Key Features

• Scalers: StandardScaler, MinMaxScaler, RobustScaler, MaxAbsScaler, QuantileTransformer, PowerTransformer.
• Encoders: OneHotEncoder, OrdinalEncoder, TargetEncoder, PolynomialFeatures, Binarizer.
• Feature Utilities: Normalizer, FunctionTransformer, MissingIndicator, discretizers, and outlier filters.
• Hardware Acceleration: SIMD, multi-threading, and optional GPU support for large tabular datasets.

Quick Start

use sklears_preprocessing::{StandardScaler, PolynomialFeatures};

let scaler = StandardScaler::default().fit(&x_train)?;
let x_scaled = scaler.transform(&x_train)?;

let poly = PolynomialFeatures::builder()
    .degree(3)
    .include_bias(false)
    .interaction_only(false)
    .build();

let x_poly = poly.fit_transform(&x_scaled)?;

Status

• Extensively covered by the 11,292 passing workspace tests in 0.1.0-beta.1.
• Provides >99% parity with scikit-learn preprocessing APIs, including sparse support.
• Future enhancements (GPU categorical encoders, streaming scalers) tracked in TODO.md.