Physics307L F09:Schedule/Week 5 agenda/CLT

From OpenWetWare

Jump to: navigation, search

Central Limit Theorem

Experimentalist point of view

Assume there is some "parent distribution" for your possible experimental measurements

We want to estimate the mean of this parent distribution (why?)

Law of large numbers

If we take a few measurements and average them: the average is the best guess for the mean of the parent population AND if we repeat this many times, our averages will have a normal distribution about the population mean

Definition involving convolutions

A pretty good experimental demo

Some other links

http://en.wikipedia.org/wiki/Illustration_of_the_central_limit_theorem

downloadable simulation: http://www.vias.org/simulations/simusoft_cenlimit.html

Java dice rolling: http://www.stat.sc.edu/~west/javahtml/CLT.html

Pretty good graphical demonstration: CLT

Personal tools