Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Define and implement a cov_to_corr() method in the CorrelationExt trait #83

Open
wants to merge 4 commits into
base: master
Choose a base branch
from
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
77 changes: 77 additions & 0 deletions src/correlation.rs
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,7 @@ use crate::errors::EmptyInput;
use ndarray::prelude::*;
use ndarray::Data;
use num_traits::{Float, FromPrimitive};
use std::error::Error;

/// Extension trait for `ArrayBase` providing functions
/// to compute different correlation measures.
Expand Down Expand Up @@ -123,6 +124,23 @@ where
A: Float + FromPrimitive;

private_decl! {}

/// Given that self is a covariance matrix, returns the appropriate correlation matrix
///
/// # Example
/// ```
/// use::ndarray::array;
/// use ndarray_stats::CorrelationExt;
///
/// let a = array![[1., 3., 5.],
/// [2., 4., 6.]];
/// let covariance = a.cov(1.).unwrap();
/// let corr = a.pearson_correlation().unwrap();
/// assert_eq!(covariance.cov_to_corr().unwrap(), corr);
/// ```
fn cov_to_corr(&self) -> Result<Array2<A>, Box<dyn Error>>
where
A: Float + FromPrimitive;
}

impl<A: 'static, S> CorrelationExt<A, S> for ArrayBase<S, Ix2>
Expand Down Expand Up @@ -179,6 +197,28 @@ where
}

private_impl! {}

fn cov_to_corr<'a>(&self) -> Result<Array2<A>, Box<dyn Error>>
where
A: Float + FromPrimitive,
{
if !self.is_square() {
return Err("A covariance matrix must be square".into());
}

let vals = self
.indexed_iter()
.map(|((x, y), v)| {
// rho_ij = sigma_ij / sqrt(sigma_ii * sigma_jj)
*v / (self[[x, x]] * self[[y, y]]).powf(A::from_f64(0.5).unwrap())
})
.collect();

match Array2::from_shape_vec(self.raw_dim(), vals) {
Ok(x) => Ok(x),
Err(e) => Err(Box::new(e)),
}
}
}

#[cfg(test)]
Expand Down Expand Up @@ -367,3 +407,40 @@ mod pearson_correlation_tests {
);
}
}

#[cfg(test)]
mod cov_to_corr_tests {
use super::*;
use ndarray::array;

#[test]
fn test_cov_2_corr_known() {
// Very basic maths that can be done in your head
let cov = array![[4., 1.], [3., 4.],];
assert_eq!(cov.cov_to_corr().unwrap(), array![[1., 0.25], [0.75, 1.]])
}

#[test]
fn test_cov_2_corr_failure() {
// A 1D array can't be a covariance matrix
let cov = array![[4., 1.],];
cov.cov_to_corr().unwrap_err();
}

#[test]
fn test_cov_2_corr_random() {
let a = array![
[0.72009497, 0.12568055, 0.55705966, 0.5959984, 0.69471457],
[0.56717131, 0.47619486, 0.21526298, 0.88915366, 0.91971245],
[0.59044195, 0.10720363, 0.76573717, 0.54693675, 0.95923036],
[0.24102952, 0.131347, 0.11118028, 0.21451351, 0.30515539],
[0.26952473, 0.93079841, 0.8080893, 0.42814155, 0.24642258]
];

// Calculating cov, and then corr should always be equivalent to calculating corr directly
let cov = a.cov(1.).unwrap();
let corr = a.pearson_correlation().unwrap();
let corr_2 = cov.cov_to_corr().unwrap();
assert!(corr.abs_diff_eq(&corr_2, 0.001));
}
}