note10 - Chapter 8 Nonparametric Bootstrap Methods 4...

This preview shows pages 1–3. Sign up to view the full content.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Chapter 8: Nonparametric Bootstrap Methods 4. Two-Sample Inference • the problem of making a confidence interval for the difference between the means of two populations using the bootstrap method • Notation Y ij = μ i + ǫ ij ∼ F i ( y ) j = 1 , 2 , . . ., n i i = 1 , 2 • t-Pivot Method Assuming Equality of Error Distributions – ǫ ij ∼ F ( ǫ ) for j = 1 , 2 , . . ., n i i = 1 , 2 ⇒ F 1 ( y ) = F 2 ( y −△ ) where △ = μ 1 − μ 2 – if F ( ǫ ) is a cdf of normal distribution, homogeneity of variances – let t = ¯ Y 1 − ¯ Y 2 − ( μ 1 − μ 2 ) S p radicalbig 1 /n 1 + 1 /n 2 where S 2 p = 2 summationdisplay i =1 n i summationdisplay j =1 ( Y ij − ¯ Y i ) 2 n 1 + n 2 − 2 ⇒ t ǫ = ¯ ǫ 1 − ¯ ǫ 2 S ǫp radicalbig 1 /n 1 + 1 /n 2 where S 2 ǫp = 2 summationdisplay i =1 n i summationdisplay j =1 ( ǫ ij − ¯ ǫ i ) 2 n 1 + n 2 − 2 – if ǫ ij ∼ N (0 , σ 2 ), − t . 975 ,df = n 1 + n 2 − 2 < − ( μ 1 − μ 2 ) S p radicalbig 1 /n 1 + 1 /n 2 < t . 975 ,df = n 1 + n 2 − 2 – t as a t-pivot for making inferences for μ 1 − μ 2 – Bootstrap interval: if the errors do not have a normal distribution (a) compute e ij = Y ij − ¯ Y i (b) from the combined observed errors, randomly select n 1 and n 2 errors with replacement for group 1 and group 2 respectively: e ij,b ⇒ compute t ǫ,b using e ij,b ’s: t e,b with b = 1 , . . ., N repetition (c) the 95% bootstrap confidence interval: ¯ Y 1 − ¯ Y 2 − t e, . 975 S p radicalbigg 1 n 1 + 1 n 2 < μ 1 − μ 2 < ¯ Y 1 − ¯ Y 2 − t e, . 025 S p radicalbigg 1 n 1 + 1 n 2 where t e, . 025 and t e, . 975 denote the 2 . 5th and 97 . 5th percentiles of the bootstrap distribution of t e,b 1 Example The mouse data (Efron and Tibshirani, 1993). Sixteen mice were randomly assigned to a treatment group or a control group. Shown are their survival times, in days, following a test surgery. Treatment 94 197 16 38 99 141 23 Control 52 104 146 10 51 30 40 27 46--------------------------------------------- Basic Statistics from the original data Treatment Control Mean 86.86 56.22 Standard Deviation 66.77 42.48 Sample Size...
View Full Document

{[ snackBarMessage ]}

Page1 / 7

note10 - Chapter 8 Nonparametric Bootstrap Methods 4...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online