statistics question

[KrazyHorse]

Originally posted by SlowwHand
Brush your hair b etor. I know that the concept he just explained sailed over your head with near g-force velocity.
Did I say I hate math yet? God, I hate math. It's like having a job where you're told every move to make. Absolutely no creativity allowed.

Just because you weren't smart enough to get to the point where creativity was what it was all about you shouldn't make generalistions

[Kuciwalker]

What's < X >?

This is the problem:

2. Consider the following situation:
The prior distribution of theta, g(theta) is Gamma(a, 1/b).
The data you have is X1 ... Xn i.i.d ~ exp(1/theta).
(i) Show that the posterior distribution of theta, g(theta | X) is also Gamma, and identify the new parameters.
(ii) What is the Bayes estimator for theta.

I could work backward and figure out what X must be, but I don't want to

[KrazyHorse]

< X > = expectation value of X

[Kuciwalker]

Ah. That couldn't be it because X is a statistic. At least, X had better be a statistic or I've got much bigger problems on my hands...

[Kuciwalker]

Completely unrelated: I am planning on taking a series of computational physics courses towards a minor in scientific computing. I have inferred that a large part of your job involves writing programs to simulate physical systems, or to do scientific calculations. How often do physicists employ people with computational physics experience but very little physics experience? I want to know how well this will improve my options on graduation.

[Jon Miller]

I seem to remember seeing this notation once and having to go ask the professor because I hadn't been paying attention in class...

Don't remember more, sorry.

Jon Miller

[KrazyHorse]

How often do physicists employ people with computational physics experience but very little physics experience?

I don't know. The larger projects (hundreds of scientists) involve dedicated software engineers building the data pipelines, but the physics knowledge they're required to have is going to be fairly minimal (I guess). Physicists aren't going to trust a compsci person to do any physics. They're going to give them every relevant equation and parameter, then look over their shoulder while the compsci does his thing.

For the problems which involve knowing a fair deal about computers and a fair deal about physics they will hire physics graduate students and phds who specialise in computational physics, not a compsci with a bit of computational physics background.

This being said, I can't imagine that having some computational physics background would hurt if you applied for a job with a big research team. You might not end up using the stuff you learned, but it might be enough to get your foot in the door.

[KrazyHorse]

Originally posted by Kuciwalker
I have inferred that a large part of your job involves writing programs to simulate physical systems, or to do scientific calculations

Not for much longer. As soon as this paper is out the door I get to stop programming. And thank god.

[KrazyHorse]

I'm beginning to be worried that my advisor is dead. He's in California and hasn't responded to my emails in the last 2 weeks.

[Addled Platypus]

bias agaisnt "Q"

start the revolt now

Q=

[b etor]

Originally posted by Kuciwalker
That's okay. I wasn't really hoping to get the answer from you anyway

btw, | = given. Thus g(theta | X = x) is the probability that theta has a particular value given that some measurement X took the value x.

Oh I know, but no one else was helping you so I figured I would.
ok.

Originally posted by SlowwHand
Brush your hair b etor. I know that the concept he just explained sailed over your head with near g-force velocity.

Brush my hair? Wtf.
Meaniepants.

[b etor]

Q= John de Lancie

[Lul Thyme]

I don't think I've ever seen that used.
I would have to review some stats stuff to even be able to make a good educated guess based on the problem you gave.
First thing I thought of was maybe a vector, since overbar is already used...