dear List:
glm(a~b+c,family=binomial,data=x)->fit
deviance(fit) returns the same as the residual deviance.
I don't not know much about logistic regression.Some book tells that:
"
Deviance (likelihood ratio statistic):
Deviance = -2log( likelihoodof the currentmodel /likelihoodof thesaturated
model)
Note:
(1). The current model is the model of interest.
(2). The saturated model is the full model that considers observed data as
parameters,
thus there are as many parameters as data points (the full model gives a
perfect fit to the data). Under this model, the maximum of the likelihood is
achieved as much as we can.
(3). If the current model is a good model, the ratio in the bracket will be
close to 1.
Otherwise, the ratio will be small.
(4). Therefore, large D suggests the current model is a poor description of
the data.
(5). The deviance for logistic regression plays the same role as the
residual sum of
squares in linear regression.
"
Is the residual deviance in fact the Deviance mentioned in that book? A got
a deviance about 2000, which looks in no sense simila to that of the
residual sum of squares in linear regression.
Thank you very much in advance!
--
View this message in context:
http://www.nabble.com/How-to-extract-the-Deviance--of-a-glm-fit-result-tf4848589.html#a13872589
Sent from the R help mailing list archive at Nabble.com.
