This preview shows pages 1–7. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: C.I. for μ from Normal Population with Unknown σ Previously assumed either: • • Now σ is unknown, but sample size not large enough for central limit the orem. Inference needs some additional assumptions. Assume N ( μ,σ 2 ) population. Interest on parameter μ but σ unknown. 1 Given random sample { X 1 ,...,X n } from N ( μ,σ 2 ) population. Previously, C.I. based on standardized score . Z = Natural to replace σ 2 with an estima tor. Use S 2 = to get a statistic T = called Studentized score . 2 Write T = The numerator is Z a standard normal. To describe the denominator first de fine the random variable W = ( n 1) S 2 σ 2 where, recall S 2 = 1 n n X i =1 ( X i ¯ X ) 2 3 The numerator of S 2 doesn’t depend on the unknown mean μ because it is subtracted out by X i ¯ X and when S 2 is divided by σ 2 , the variance σ 2 can cels out in numerator and denomina tor. Therefore W is dimensionless ran dom variable. It can be shown that the distribution of W has a chisquare distribution which is a family that depends only on one parameter, called the degrees of freedom , and given in this case by r = n 1 . 4 Simulated chi square 10000 times with r = 9 and superimposed pdf (red curve). 5 10 15 20 25 30 35 40 45 50 500 1000 1500 2000 2500 3000 3500 4000 5 Matlab code for simulation and fit . Simulate chisquare on 9 degrees of freedom from nor mal samples of size 10....
View
Full
Document
This note was uploaded on 01/03/2012 for the course EE 1244 taught by Professor Drera during the Fall '10 term at Conestoga.
 Fall '10
 drera

Click to edit the document details