We generate synthetic data from:

y = 2·x + 1 + Normal(0, σ²)

and then fit a linear regression model:

ŷ = a·x + b

so you can compare the recovered parameters (a, b) with the true values (2, 1).