Fix covariance and add tests

src/covariance.rs

@@ -4,8 +4,24 @@ use serde_derive::{Deserialize, Serialize};
 
 use crate::Merge;
 
-/// Estimate the arithmetic mean and the covariance of a sequence of number pairs
+/// Estimate the arithmetic means and the covariance of a sequence of number pairs
 /// ("population").
+///
+/// Because the variances are calculated as well, this can be used to calculate the Pearson
+/// correlation coefficient.
+///
+///
+/// ## Example
+///
+/// ```
+/// use average::Covariance;
+///
+/// let a: Covariance = [(1., 5.), (2., 4.), (3., 3.), (4., 2.), (5., 1.)].iter().cloned().collect();
+/// assert_eq!(a.mean_x(), 3.);
+/// assert_eq!(a.mean_y(), 3.);
+/// assert_eq!(a.population_covariance(), -2.5);
+/// assert_eq!(a.sample_covariance(), -2.0);
+/// ```
 #[derive(Debug, Clone)]
 #[cfg_attr(feature = "serde1", derive(Serialize, Deserialize))]
 pub struct Covariance {
@@ -40,13 +56,13 @@ impl Covariance {
         let delta_x = x - self.avg_x;
         let delta_y = y - self.avg_y;
 
-        self.avg_x += delta_x / self.n.to_f64().unwrap();
+        self.avg_x += delta_x / n;
         self.sum_x_2 += delta_x * delta_x * n * (n - 1.);
 
-        self.avg_y += delta_y / self.n.to_f64().unwrap();
+        self.avg_y += delta_y / n;
         self.sum_y_2 += delta_y * delta_y * n * (n - 1.);
 
-        self.sum_prod += delta_x * delta_y;
+        self.sum_prod += delta_x * (y - self.avg_y);
     }
 
     /// Calculate the population covariance of the sample.
@@ -56,7 +72,7 @@ impl Covariance {
     /// Returns NaN for an empty sample.
     #[inline]
     pub fn population_covariance(&self) -> f64 {
-        if self.n < 2 {
+        if self.n < 1 {
             return f64::NAN;
         }
         self.sum_prod / self.n.to_f64().unwrap()

src/lib.rs

@@ -93,6 +93,7 @@
 #![forbid(missing_docs)]
 #![forbid(missing_debug_implementations)]
 #![cfg_attr(feature = "nightly", feature(generic_const_exprs))]
+#[cfg(feature = "std")] extern crate std;
 
 #[macro_use]
 mod macros;

tests/integration/covariance.rs

@@ -0,0 +1,43 @@
+use average::Covariance;
+
+#[test]
+fn simple() {
+    let mut cov = Covariance::new();
+    assert!(cov.mean_x().is_nan());
+    assert!(cov.mean_y().is_nan());
+    assert!(cov.population_covariance().is_nan());
+    assert!(cov.sample_covariance().is_nan());
+    assert!(cov.population_pearson().is_nan());
+    assert!(cov.sample_pearson().is_nan());
+
+    cov.add(1., 5.);
+    assert_eq!(cov.mean_x(), 1.);
+    assert_eq!(cov.mean_y(), 5.);
+    assert_eq!(cov.population_covariance(), 0.);
+    assert!(cov.sample_covariance().is_nan());
+    // TODO: pearson
+
+    cov.add(2., 4.);
+    assert_eq!(cov.mean_x(), 1.5);
+    assert_eq!(cov.mean_y(), 4.5);
+    assert_eq!(cov.population_covariance(), -0.25);
+    assert_eq!(cov.sample_covariance(), -0.5);
+
+    cov.add(3., 3.);
+    assert_eq!(cov.mean_x(), 2.);
+    assert_eq!(cov.mean_y(), 4.);
+    assert_eq!(cov.population_covariance(), -2./3.);
+    assert_eq!(cov.sample_covariance(), -1.);
+
+    cov.add(4., 2.);
+    assert_eq!(cov.mean_x(), 2.5);
+    assert_eq!(cov.mean_y(), 3.5);
+    assert_eq!(cov.population_covariance(), -1.25);
+    assert_eq!(cov.sample_covariance(), -5./3.);
+
+    cov.add(5., 1.);
+    assert_eq!(cov.mean_x(), 3.);
+    assert_eq!(cov.mean_y(), 3.);
+    assert_eq!(cov.population_covariance(), -2.0);
+    assert_eq!(cov.sample_covariance(), -2.5);
+}

tests/integration/main.rs

@@ -22,3 +22,4 @@ mod skewness;
 #[cfg(feature = "std")]
 mod streaming_stats;
 mod weighted_mean;
+mod covariance;