Tell me about the whole gaussian vs mandelbrotian thing and how the bell curve is an useless piece of shit

Tell me about the whole gaussian vs mandelbrotian thing and how the bell curve is an useless piece of shit.

>an useless
>an hero
you can't have one without the other :^)

Is that a roundabout way of saying you don't know what I'm asking about?

No, this is a roundabout way of saying that your incapability to spell things correctly qualifies yourself for suicide.

I'm not from usa and was taught that it's an if it's preceding a vowel.

Because "useless" is pronounced "yoos-less", it should be preceded by a vowel.
For example, one should say "an honest man", not "a honest man", because "honest" has a silent h. The reverse is true for useless - it's spelled with a vowel but it's pronounced with a consonant.

Just to add on to this, when foreigners first learn English, they are infact taught that 'an' precedes a vowel, and 'a' precedes a consonant. It is the easiest way of explaining it and in most cases it is true. However, the factor you need to take into consideration is not how you spell the word, but how you pronounce it.

Ignore him, he's just an asshole

>should be


>muh objective moralism
look at this 20 yo fagget

>muh relativism
look at this kike

>your incapability to spell things correctly qualifies yourself

>Because "useless" is pronounced "yoos-less", it should be preceded by a vowel.

>uses another vowel-word as his example.

Because "useless" is pronounced jouseless, it should be preceded by a vowel.

If the letter is silent (often h words), go by the second letter.

if its a vowel: an

if its not a vowel: a

"an herb" because its really "an erb"

"a useless" because u is not silent

>if its a vowel: an
>"a useless" because u is not silent

Gaussian is useless because statisticians try to take nonparametric data and fit it to a normal distribution. That's about 50% of their job: fitting data to a normal distirbution.

>"an herb" because its really "an erb"
Except in England people pronounce "herb" correctly, so it's "a herb"
;^)

What are the best methods of fitting data?
Linear regression is shit.
Quadratic spline maybe is good?
Why do statisticians always try to fit data to normal distributions?

Unless you're from Yorkshire.

They're nice to work with.

You can use non-paremetric regression. There is also logistic regression, probit regresion, ridge regression. And even if some of the normal assumptions are broken, there are ways to help correct that. Also, I think the Gaussian curve is used in physics.

Doesn't that make them frauds?

Srs question, does anything more complicated than multiple regression work irl?

>Also, I think the Gaussian curve is used in physics.
I've used it for optics-related stuff (Gaussian-profile beams, some spectral characteristics), but I don't think it's related anyhow to its use in statistics. Well, maybe except for the tendency to put it where it doesn't truly belong, just for its nice properties (being its own Fourier transform, mostly).