#quadratic-discriminant-analysis

Question 1: Single choice

Given $Y \in \{0, 1\}$ , the input is $X \mid Y \sim \mathcal{N}(\mu_Y, \Sigma_Y)$ in $\mathbb{R}^d$ , with $\mu_0 \ne \mu_1$ and two different positive definite covariance matrices $\Sigma_0 \ne \Sigma_1$ . What does the boundary between $\{h^* = 0\}$ and $\{h^* = 1\}$ of the Bayes classifier look like in general?

A hyperplane, as when the covariances are equal1
Always an ellipse (an ellipsoid) around one of the means2
A quadric, e.g. a hyperbola in the plane3
The perpendicular bisector of $\mu_0$ and $\mu_1$ 4

#quadratic-discriminant-analysis

Timed test