examples: simple linear regression

pull/3615/head
kemalcan bora 2020-01-31 04:44:38 +03:00 committed by GitHub
parent de832375b7
commit c90b53b89d
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 56 additions and 0 deletions

View File

@ -0,0 +1 @@
simple_linear_regression

View File

@ -0,0 +1,55 @@
import math
struct LinearResult {
r2 f64
intercept f64
slope f64
dependent_variable_means f64
independent_variable_means f64
}
fn linearrelationship(independent_variable []int, dependent_variable []int) LinearResult {
// Objective :
// Find what is the linear relationship between two dataset X and Y?
// x := independent variable
// y := dependent variable
mut sum_r2_x := 0
mut sum_r2_y := 0
mut sum_xy := 0
mut sum_x := 0
mut sum_y := 0
for i in independent_variable {
sum_x += i
sum_r2_x += i * i
}
for yi in dependent_variable {
sum_y += yi
sum_r2_y += yi * yi
}
x_means := sum_x / independent_variable.len
y_means := sum_y / dependent_variable.len
for index, x_value in independent_variable {
sum_xy += x_value * dependent_variable[index]
}
// /Slope = (∑y)(∑x²) - (∑x)(∑xy) / n(∑x²) - (∑x)²
slope_value := f64((sum_y * sum_r2_x) - (sum_x * sum_xy)) / f64((sum_r2_x * independent_variable.len) - (sum_x * sum_x))
// /Intercept = n(∑xy) - (∑x)(∑y) / n(∑x²) - (∑x)²
intercept_value := f64((independent_variable.len * sum_xy) - (sum_x * sum_y)) / f64((independent_variable.len * sum_r2_x) - (sum_x * sum_x))
// Regression equation = Intercept + Slope x
// R2 = n(∑xy) - (∑x)(∑y) / sqrt([n(∑x²)-(∑x)²][n(∑y²)-(∑y)²]
r2_value := f64((independent_variable.len * sum_xy) - (sum_x * sum_y)) / math.sqrt(f64((sum_r2_x * independent_variable.len) - (sum_x * sum_x)) * f64((sum_r2_y * dependent_variable.len) - (sum_y * sum_y)))
return LinearResult{
r2: r2_value
intercept: intercept_value
slope: slope_value
independent_variable_means: x_means
dependent_variable_means: y_means
}
}
fn main() {
independent_variable := [4, 5, 6, 7, 10]
dependent_variable := [3, 8, 20, 30, 12]
result := linearrelationship(independent_variable, dependent_variable)
println(result)
}