Regression Coefficients for Uniform Points in a Triangle

Reported in interviews at Squarepoint Capital

Problem statement

Points $(x_i, y_i)$ are uniformly distributed within the triangular region defined by $0 < x \leq 1$, $0 < y \leq 1$, and $x + y \geq 1.5$.

A linear regression model $\hat{y} = bx + k$ is fit to these points.

What are the expected values $E[b]$ and $E[k]$?

Hint: Use the fact that for the population regression, $\hat{y}(x) = E[Y \mid X = x]$, and that the conditional distribution $p(y \mid x)$ is uniform over $[1.5 - x,\, 1]$ for $x \in [0.5, 1]$.

Show hint

Before computing anything, sketch the region: where does it actually live in the unit square?

Relevant roles: Data Scientist, Applied Scientist and ML Engineer.

Related free problems