9 releases (5 breaking)

0.7.0 Oct 16, 2023
0.6.1 Dec 3, 2022
0.6.0 Jun 15, 2022
0.5.1 Mar 1, 2022
0.2.1 Nov 29, 2020

#586 in Machine learning

Download history 667/week @ 2024-01-25 846/week @ 2024-02-01 732/week @ 2024-02-08 796/week @ 2024-02-15 638/week @ 2024-02-22 579/week @ 2024-02-29 550/week @ 2024-03-07 531/week @ 2024-03-14 526/week @ 2024-03-21 459/week @ 2024-03-28 586/week @ 2024-04-04 711/week @ 2024-04-11 606/week @ 2024-04-18 521/week @ 2024-04-25 565/week @ 2024-05-02 421/week @ 2024-05-09

2,239 downloads per month
Used in 4 crates (3 directly)

MIT/Apache

290KB
5.5K SLoC

Kernel methods

linfa-kernel provides methods for dimensionality expansion.

The Big Picture

linfa-kernel is a crate in the linfa ecosystem, an effort to create a toolkit for classical Machine Learning implemented in pure Rust, akin to Python's scikit-learn.

In machine learning, kernel methods are a class of algorithms for pattern analysis, whose best known member is the support vector machine. They owe their name to the kernel functions, which maps the features to some higher-dimensional target space. Common examples for kernel functions are the radial basis function (euclidean distance) or polynomial kernels.

Current State

linfa-kernel currently provides an implementation of kernel methods for RBF and polynomial kernels, with sparse or dense representation. Further a k-neighbour approximation allows to reduce the kernel matrix size.

Low-rank kernel approximation are currently missing, but are on the roadmap. Examples for these are the Nyström approximation or Quasi Random Fourier Features.

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.

Dependencies

~5MB
~94K SLoC