Skip to content

Comment on Ideas in statistics that have powered AIparent

Comments

Given a Dirac comb
c(x) = 1/n Σ_i=1^n δ(х - x_i)

Sorry for the pedantry, but a mixture of Dirac distributions is almost always not a Dirac comb. Notice that a mixture of Dirac distributions is a Dirac comb only if you have an infinite number of equally-separated samples (and empirical distributions tend to have a finite number of samples).

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.