Gaussian processes

egobox-gp provides a Rust implementation of gaussian process regression. It is a Rust port of some kriging algorithms of the SMT Python library.

The big picture

egobox-gp is a library crate in the top-level package egobox.

Current state

egobox-gp currently provides a Gaussian Process implementation with the following features:

• Regression model choice: constant, linear or quadratic
• Correlation model (kernel) choice: squared exponential, absolute exponential, matern 3/2, matern 5/2
• Handling of high dimensional problem using PLS (linfa-pls)

Examples

There is an usage example in the examples/ directory. To run, use:

$ cargo run --release --example kriging --features linfa/intel-mkl-static

License

Licensed under the Apache License, Version 2.0 http://www.apache.org/licenses/LICENSE-2.0