Wednesday, November 14, 2007

MCMC Hammer

After working on some personal research related to music, I think I have found a way to better explain the MCMC method. Below is a digram I have made that might just revolutionize the world of music and statistics:
I have so much faith in this that I have opened up a CafePress store for it. Check it out.


(Thanks Nicole for the idea! More about the MCMC method, more about MC Hammer.)

8 comments:

The Math Ninja said...

I'm excited!!

Braveharte said...

How are monte carlo and bootstrapping different?

Think Frustrated said...

I'm going to get the thong for myself.

I tutor math, but this is way out of my league. After reading the wikipedia article three times, my nose started to bleed.

Ookami Snow said...

@julia - Now I am not an expert in such things, but i think MCMC is used to predict missing values in a data set, where as bootstrapping is trying to get parameters from the data for hypothesis testing (such as sample mean and variance.)

@think- Is your nose supposed to bleed form intense thought? You might want to get that checked out, it still might be under warranty.

Think Frustrated said...

I think my brain was just rejecting all that talk about algorithms. It tried to understand, but fell over and scraped itself, and started to bleed. I'm sure for someone like you this is the equivelant of disucussing the different between a direct object and a predicate nominative, and their effect on overall sentence structure and effect. Or maybe not.

Tuba Queen said...

I like the MC Hammer part? No pretending here, I have no idea what you are talking about. But I am still thinking of buying a piece of clothing with it on there. Just to look smart.

The Math Ninja said...

In bootstrapping you resample from the sample you have. In MCMC method, you are using the data you are given to generate data that follows the same distribution as that data. Then after you've converged, you are taking more "samples" from the same distribution.

Spinning Girl said...

That is hysterical!