darrenbaldwin07 / linfa

A Rust machine learning framework.

Linfa

linfa (Italian) / sap (English):

The vital circulating fluid of a plant.

linfa aims to provide a comprehensive toolkit to build Machine Learning applications with Rust.

Kin in spirit to Python's scikit-learn, it focuses on common preprocessing tasks and classical ML algorithms for your everyday ML tasks.

Website | Community chat

Current state

Where does linfa stand right now? Are we learning yet?

linfa currently provides sub-packages with the following algorithms:

Name	Purpose	Status	Category	Notes
clustering	Data clustering	Tested / Benchmarked	Unsupervised learning	Clustering of unlabeled data; contains K-Means, Gaussian-Mixture-Model, DBSCAN and OPTICS
kernel	Kernel methods for data transformation	Tested	Pre-processing	Maps feature vector into higher-dimensional space
linear	Linear regression	Tested	Partial fit	Contains Ordinary Least Squares (OLS), Generalized Linear Models (GLM)
elasticnet	Elastic Net	Tested	Supervised learning	Linear regression with elastic net constraints
logistic	Logistic regression	Tested	Partial fit	Builds two-class logistic regression models
reduction	Dimensionality reduction	Tested	Pre-processing	Diffusion mapping and Principal Component Analysis (PCA)
trees	Decision trees	Tested / Benchmarked	Supervised learning	Linear decision trees
svm	Support Vector Machines	Tested	Supervised learning	Classification or regression analysis of labeled datasets
hierarchical	Agglomerative hierarchical clustering	Tested	Unsupervised learning	Cluster and build hierarchy of clusters
bayes	Naive Bayes	Tested	Supervised learning	Contains Gaussian Naive Bayes
ica	Independent component analysis	Tested	Unsupervised learning	Contains FastICA implementation
pls	Partial Least Squares	Tested	Supervised learning	Contains PLS estimators for dimensionality reduction and regression
tsne	Dimensionality reduction	Tested	Unsupervised learning	Contains exact solution and Barnes-Hut approximation t-SNE
preprocessing	Normalization & Vectorization	Tested / Benchmarked	Pre-processing	Contains data normalization/whitening and count vectorization/tf-idf
nn	Nearest Neighbours & Distances	Tested / Benchmarked	Pre-processing	Spatial index structures and distance functions
ftrl	Follow The Regularized Leader - proximal	Tested / Benchmarked	Partial fit	Contains L1 and L2 regularization. Possible incremental update

We believe that only a significant community effort can nurture, build, and sustain a machine learning ecosystem in Rust - there is no other way forward.

If this strikes a chord with you, please take a look at the roadmap and get involved!

BLAS/Lapack backend

Some algorithm crates need to use an external library for linear algebra routines. By default, we use a pure-Rust implementation. However, you can also choose an external BLAS/LAPACK backend library instead, by enabling the blas feature and a feature corresponding to your BLAS backend. Currently you can choose between the following BLAS/LAPACK backends: openblas, netblas or intel-mkl.

Backend	Linux	Windows	macOS
OpenBLAS	✔️	-	-
Netlib	✔️	-	-
Intel MKL	✔️	✔️	✔️

Each BLAS backend has two features available. The feature allows you to choose between linking the BLAS library in your system or statically building the library. For example, the features for the intel-mkl backend are intel-mkl-static and intel-mkl-system.

An example set of Cargo flags for enabling the Intel MKL backend on an algorithm crate is --features blas,linfa/intel-mkl-system. Note that the BLAS backend features are defined on the linfa crate, and should only be specified for the final executable.

License

Dual-licensed to be compatible with the Rust project.

Licensed under the Apache License, Version 2.0 http://www.apache.org/licenses/LICENSE-2.0 or the MIT license http://opensource.org/licenses/MIT, at your option. This file may not be copied, modified, or distributed except according to those terms.