Deviance for Binary Data Notes

Unm r computes the null deviance for all i g i 0

Info iconThis preview shows page 1. Sign up to view the full content.

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

Unformatted text preview: ; Y ) − l (b; Y )] N =2 yi log i =1 yi ˆ yi + (ni − yi )log ni − yi ˆ ni − yi D s approximate sampling distribution is chi-square. UNM R computes the null deviance. For all i ˆ g (πi ) = β0 → π = g −1 (β0 ) ˆ ˆ Fitted values yi = ni π ˆ Null deviance: N −2 yi log (ˆ ) + (ni − yi )log (1 − π ) + π ˆ i =1 ni yi with degrees of freedom N − 1. Residual deviance: g (πi ) = β0 + β1 Xi 1 + . . . + βp Xpi ˆ πi are estimated with covariates and yi = ni πi ˆ ˆ UNM Residual deviance: N D = −2 yi log (ˆi ) + (ni − yi )log (1 − πi ) + π ˆ i =1 ni yi with degrees of freedom N − (p + 1). Hypothesis testing: H0 : β1 = β2 = . . . = βp = 0 vs. H1 : at least one βi = 0. D ∗ = Null deviance − Residual deviance dof = (N − 1) − (N − (p + 1)) = p. D ∗ approximately follows a χ2p) . ( UNM Alternatively if we just focus on the Residual deviance (grouped Binomial case), dof = (n1 + n2 + . . . + nN ) − (p + 1) or dof = no....
View Full Document

This note was uploaded on 01/27/2014 for the course STAT 574 taught by Professor Gabrielhuerta during the Fall '13 term at New Mexico.

Ask a homework question - tutors are online