random-world

This is a rust implementation of Machine Learning (ML) methods for confident prediction (e.g., Conformal Predictors) and related ones introduced in the book Algorithmic Learning in a Random World (ALRW).

Goals

These are the main goals of this library. The fact that something appears here does not imply that it has already been fulfilled.

• Fast implementation of methods described in ALRW book
• Be well documented
• CP method should be able to use scoring classifiers from existing ML libraries (e.g., rusty-machine, rustlearn)
• Easily interface with other languages (e.g., Python)
• Provide standalone programs for some functions (e.g., in the way libsvm does for SVM methods)

Install

Using cargo, as soon as this code is published as a crate on crates.io.

Quick Intro

Using a deterministic (i.e., non smooth) Conformal Predictor with k-NN nonconformity measure (k=2) and significance level epsilon=0.3. The prediction region will contain the correct label with probability 1-epsilon.

```rust

[macro_use(array)]

extern crate ndarray; extern crate random_world;

use randomworld::cp::*; use randomworld::ncm::*;

let ncm = KNN::new(2); let mut cp = CP::new(ncm, Some(0.3)); let traininputs = array![[0., 0.], [1., 0.], [0., 1.], [1., 1.], [2., 2.], [1., 2.]]; let traintargets = array![0, 0, 0, 1, 1, 1]; let test_inputs = array![[2., 1.], [2., 2.]];

// Train and predict cp.train(&traininputs.view(), &traintargets.view()) .expect("Failed prediction"); let preds = cp.predict(&test_inputs.view()) .expect("Failed to predict"); assert!(preds == array![[false, true], [false, true]]); ``` More examples on deterministic/smooth Conformal Predictors at CP.

Features

Methods: - [x] Deterministic and smoothed Conformal Predictors (aka, transductive CP) - [ ] Deterministic and smoothed Inductive Conformal Predictors (ICP) - [ ] Plug-in martingales for exchangeability testing - [ ] Venn Predictors

Nonconformity measures: - [x] k-NN - [ ] KDE - [ ] Generic wrapper around existing ML scorers (e.g., rusty-machine)

Bindings: - [ ] Python bindings

Binaries: - [ ] CP - [ ] Martingales

Authors

• Giovanni Cherubin giocher.com

Similar Projects

• nonconformist is a well maintained Python implementation of CP and ICP.