petal-decomposition

petal-decomposition provides matrix decomposition algorithms including PCA (principal component analysis) and ICA (independent component analysis).

Requirements

• Rust ≥ 1.51
• BLAS/LAPACK backend (OpenBLAS, Netlib, or Intel MKL)

Features

• PCA with exact, full SVD (singular value decomposition)
• PCA with randomized, truncated SVD
• FastICA

Crate Features

• Use one of intel-mkl-static, intel-mkl-system, netlib-static, netlib-system, openblas-static, and openblas-system to select a BLAS/LAPACK backend. See ndarray-linalg's documentation for details.
• serialization enables serialization/deserialization using serde.

Examples

The following example shows how to apply PCA to an array of three samples, and obtain singular values as well as how much variance each component explains.

```rust use ndarray::arr2; use petal_decomposition::Pca;

let x = arr2(&[[0f64, 0f64], [1f64, 1f64], [2f64, 2f64]]); let mut pca = PcaBuilder::new(2).build(); // Keep two dimensions. pca.fit(&x).unwrap();

let s = pca.singularvalues(); // [2f64, 0f64] let v = pca.explainedvarianceratio(); // [1f64, 0f64] let y = pca.transform(&x).unwrap(); // [-2f64.sqrt(), 0f64, 2f64.sqrt()] ```

License

Copyright 2020-2021 Petabi, Inc.

Licensed under Apache License, Version 2.0 (the "License"); you may not use this crate except in compliance with the License.

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See LICENSE for the specific language governing permissions and limitations under the License.

Contribution

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be licensed as above, without any additional terms or conditions.