work/smartcore
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge potential next release v0.4 (#187) Breaking Changes

Browse Source 
* First draft of the new n-dimensional arrays + NB use case
* Improves default implementation of multiple Array methods
* Refactors tree methods
* Adds matrix decomposition routines
* Adds matrix decomposition methods to ndarray and nalgebra bindings
* Refactoring + linear regression now uses array2
* Ridge & Linear regression
* LBFGS optimizer & logistic regression
* LBFGS optimizer & logistic regression
* Changes linear methods, metrics and model selection methods to new n-dimensional arrays
* Switches KNN and clustering algorithms to new n-d array layer
* Refactors distance metrics
* Optimizes knn and clustering methods
* Refactors metrics module
* Switches decomposition methods to n-dimensional arrays
* Linalg refactoring - cleanup rng merge (#172)
* Remove legacy DenseMatrix and BaseMatrix implementation. Port the new Number, FloatNumber and Array implementation into module structure.
* Exclude AUC metrics. Needs reimplementation
* Improve developers walkthrough

New traits system in place at `src/numbers` and `src/linalg`
Co-authored-by: Lorenzo <tunedconsulting@gmail.com>

* Provide SupervisedEstimator with a constructor to avoid explicit dynamical box allocation in 'cross_validate' and 'cross_validate_predict' as required by the use of 'dyn' as per Rust 2021
* Implement getters to use as_ref() in src/neighbors
* Implement getters to use as_ref() in src/naive_bayes
* Implement getters to use as_ref() in src/linear
* Add Clone to src/naive_bayes
* Change signature for cross_validate and other model_selection functions to abide to use of dyn in Rust 2021
* Implement ndarray-bindings. Remove FloatNumber from implementations
* Drop nalgebra-bindings support (as decided in conf-call to go for ndarray)
* Remove benches. Benches will have their own repo at smartcore-benches
* Implement SVC
* Implement SVC serialization. Move search parameters in dedicated module
* Implement SVR. Definitely too slow
* Fix compilation issues for wasm (#202)

Co-authored-by: Luis Moreno <morenol@users.noreply.github.com>
* Fix tests (#203)

* Port linalg/traits/stats.rs
* Improve methods naming
* Improve Display for DenseMatrix

Co-authored-by: Montana Low <montanalow@users.noreply.github.com>
Co-authored-by: VolodymyrOrlov <volodymyr.orlov@gmail.com>
This commit is contained in:
Lorenzo
2022-10-31 10:44:57 +00:00
committed by morenol
parent a32eb66a6a
commit a7fa0585eb
110 changed files with 10327 additions and 9107 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/metrics/auc.rs
+47 -24
View File
@@ -7,11 +7,12 @@
//! Example:
//! ```
//! use smartcore::metrics::auc::AUC;
//! use smartcore::metrics::Metrics;
//!
//! let y_true: Vec<f64> = vec![0., 0., 1., 1.];
//! let y_pred: Vec<f64> = vec![0.1, 0.4, 0.35, 0.8];
//!
//! let score1: f64 = AUC {}.get_score(&y_true, &y_pred);
//! let score1: f64 = AUC::new().get_score(&y_true, &y_pred);
//! ```
//!
//! ## References:
@@ -20,32 +21,52 @@
//! * ["The ROC-AUC and the Mann-Whitney U-test", Haupt, J.](https://johaupt.github.io/roc-auc/model%20evaluation/Area_under_ROC_curve.html)
#![allow(non_snake_case)]
use std::marker::PhantomData;
#[cfg(feature = "serde")]
use serde::{Deserialize, Serialize};
use crate::algorithm::sort::quick_sort::QuickArgSort;
use crate::linalg::BaseVector;
use crate::math::num::RealNumber;
use crate::linalg::basic::arrays::{Array1, ArrayView1, MutArrayView1};
use crate::numbers::basenum::Number;
use crate::metrics::Metrics;
/// Area Under the Receiver Operating Characteristic Curve (ROC AUC)
#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
#[derive(Debug)]
pub struct AUC {}
pub struct AUC<T> {
_phantom: PhantomData<T>,
}
impl AUC {
impl<T: Number + Ord> Metrics<T> for AUC<T> {
/// create a typed object to call AUC functions
fn new() -> Self {
Self {
_phantom: PhantomData,
}
}
fn new_with(_parameter: T) -> Self {
Self {
_phantom: PhantomData,
}
}
/// AUC score.
/// * `y_true` - cround truth (correct) labels.
/// * `y_pred_probabilities` - probability estimates, as returned by a classifier.
pub fn get_score<T: RealNumber, V: BaseVector<T>>(&self, y_true: &V, y_pred_prob: &V) -> T {
/// * `y_true` - ground truth (correct) labels.
/// * `y_pred_prob` - probability estimates, as returned by a classifier.
fn get_score(
&self,
y_true: &dyn ArrayView1<T>,
y_pred_prob: &dyn ArrayView1<T>,
) -> f64 {
let mut pos = T::zero();
let mut neg = T::zero();
let n = y_true.len();
let n = y_true.shape();
for i in 0..n {
if y_true.get(i) == T::zero() {
if y_true.get(i) == &T::zero() {
neg += T::one();
} else if y_true.get(i) == T::one() {
} else if y_true.get(i) == &T::one() {
pos += T::one();
} else {
panic!(
@@ -55,21 +76,21 @@ impl AUC {
}
}
let mut y_pred = y_pred_prob.to_vec();
let y_pred = y_pred_prob.clone();
let label_idx = y_pred.quick_argsort_mut();
let label_idx = y_pred.argsort();
let mut rank = vec![T::zero(); n];
let mut rank = vec![0f64; n];
let mut i = 0;
while i < n {
if i == n - 1 || y_pred[i] != y_pred[i + 1] {
rank[i] = T::from_usize(i + 1).unwrap();
if i == n - 1 || y_pred.get(i) != y_pred.get(i + 1) {
rank[i] = (i + 1) as f64;
} else {
let mut j = i + 1;
while j < n && y_pred[j] == y_pred[i] {
while j < n && y_pred.get(j) == y_pred.get(i) {
j += 1;
}
let r = T::from_usize(i + 1 + j).unwrap() / T::two();
let r = (i + 1 + j) as f64 / 2f64;
for rank_k in rank.iter_mut().take(j).skip(i) {
*rank_k = r;
}
@@ -78,14 +99,16 @@ impl AUC {
i += 1;
}
let mut auc = T::zero();
let mut auc = 0f64;
for i in 0..n {
if y_true.get(label_idx[i]) == T::one() {
if y_true.get(label_idx[i]) == &T::one() {
auc += rank[i];
}
}
let pos = pos.to_f64().unwrap();
let neg = neg.to_f64().unwrap();
(auc - (pos * (pos + T::one()) / T::two())) / (pos * neg)
T::from(auc - (pos * (pos + 1f64) / 2.0)).unwrap() / T::from(pos * neg).unwrap()
}
}
@@ -99,8 +122,8 @@ mod tests {
let y_true: Vec<f64> = vec![0., 0., 1., 1.];
let y_pred: Vec<f64> = vec![0.1, 0.4, 0.35, 0.8];
let score1: f64 = AUC {}.get_score(&y_true, &y_pred);
let score2: f64 = AUC {}.get_score(&y_true, &y_true);
let score1: f64 = AUC::new().get_score(&y_true, &y_pred);
let score2: f64 = AUC::new().get_score(&y_true, &y_true);
assert!((score1 - 0.75).abs() < 1e-8);
assert!((score2 - 1.0).abs() < 1e-8);