Bayes decision explorer
Class 0: p(x | Y=0)
mean μ₀
spread σ₀
Class 1: p(x | Y=1)
mean μ₁
spread σ₁
Prior and costs
prior π₁ = P(Y=1)
ℓ₀₁ (miss a 1)
ℓ₁₀ (false 1)
Shape of the class conditionals
shape
Gaussian
triangle
uniform (box)
Triangle and box: the density is non-zero on [μ − w, μ + w].
π₀ℓ₁₀ p(x|0)
π₁ℓ₀₁ p(x|1)
expected loss of the Bayes rule
Bayes boundary
ML boundary
MAP boundary