Formula
y = a + bx
With least-squares, b = Cov(x, y) / Var(x), and a = ȳ − b x̄.
Example
For points (1, 2), (2, 4), (3, 5), the fitted line is approximately y = 1 + 1.5x.
FAQs
What does regression show?
It models the relationship between an independent variable and a dependent variable.
Is this always exact?
It provides the best linear fit for the data under least-squares assumptions.