linreg-rs

linreg is a small crate that calculates linear regressions. It works without stdlib or memory allocation and has few dependencies. Example:

let xs: Vec<f64> = vec![1.0, 2.0, 3.0, 4.0, 5.0];
let ys: Vec<f64> = vec![2.0, 4.0, 5.0, 4.0, 5.0];

assert_eq!(Some((0.6, 2.2)), linear_regression(&xs, &ys));

It supports tuples, separate vectors for x and y values and template floating point types.

lib.rs:

Linear regression

linreg calculates linear regressions for two dimensional measurements, also known as simple linear regression.

Base for all calculations of linear regression is the simple model found in https://en.wikipedia.org/wiki/Ordinary_least_squares#Simple_linear_regression_model.

Example use

   use linreg::{linear_regression, linear_regression_of};

   // Example 1: x and y values stored in two different vectors
   let xs: Vec<f64> = vec![1.0, 2.0, 3.0, 4.0, 5.0];
   let ys: Vec<f64> = vec![2.0, 4.0, 5.0, 4.0, 5.0];

   assert_eq!(Ok((0.6, 2.2)), linear_regression(&xs, &ys));


   // Example 2: x and y values stored as tuples
   let tuples: Vec<(f32, f32)> = vec![(1.0, 2.0),
                                      (2.0, 4.0),
                                      (3.0, 5.0),
                                      (4.0, 4.0),
                                      (5.0, 5.0)];

   assert_eq!(Ok((0.6, 2.2)), linear_regression_of(&tuples));


   // Example 3: directly operating on integer (converted to float as required)
   let xs: Vec<u8> = vec![1, 2, 3, 4, 5];
   let ys: Vec<u8> = vec![2, 4, 5, 4, 5];

   assert_eq!(Ok((0.6, 2.2)), linear_regression(&xs, &ys));