[R] What is the "deviance " and the "change in deviance" of a logistic regression?

Bin Yue leffgh at 163.com
Sun Nov 25 13:30:32 CET 2007


Dear all:
     I am confounded by the "deviance" and the "change in deviance" of a
logistic regression ?
     I have a object like this:
   summary(glm1)

Call:
glm(formula = status ~ no.filter + light.filter_2, family = binomial, 
    data = re)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-0.9774  -0.8063  -0.7586   1.4864   3.0635  

Coefficients:
                Estimate Std. Error z value Pr(>|z|)    
(Intercept)    -1.155042   0.087014 -13.274  < 2e-16 ***
no.filter       0.057531   0.016475   3.492 0.000479 ***
light.filter_2 -0.037554   0.008619  -4.357 1.32e-05 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 3655.4  on 3172  degrees of freedom
Residual deviance: 3612.5  on 3170  degrees of freedom
  (362 observations deleted due to missingness)
AIC: 3618.5

Number of Fisher Scoring iterations: 5

> deviance(glm1)
[1] 3612.516

is the "deviance(glm1)" the "deviance" ?
If yes, what is the "change in deviance"? Is it "residual
deviance-null.deviance"?
Thanks all .

Best wishes,
Bin Yue 

-----
Best regards,
Bin Yue

*************
student for a Master program in South Botanical Garden , CAS

-- 
View this message in context: http://www.nabble.com/%EF%BC%B7hat-is-the-%22deviance-%22--and-the-%22change-in-deviance%22-of-a-logistic-regression--tf4870005.html#a13935092
Sent from the R help mailing list archive at Nabble.com.



More information about the R-help mailing list