i am using vglm for multiple logistic regression.
i have 1 response variable (total 4 category)
and 5 predictor.
Call:
vglm(formula = class ~ PC1 + PC2 + PC3 + PC4 + PC5, family = multinomial(),
na.action = na.pass)
Coefficients:
(Intercept):1 (Intercept):2 PC1:1 PC1:2 PC2:1
-0.5480417 -1.0716498 0.5146799 0.1578941 -0.3111874
PC2:2 PC3:1 PC3:2 PC4:1 PC4:2
0.5213314 -0.9584294 -0.9889684 0.8510812 1.2110904
PC5:1 PC5:2
0.5832257 0.5126038
Degrees of Freedom: 330 Total; 318 Residual
Residual Deviance: 216.9244
Log-likelihood: -108.4622
i am not understanding whether this model is good or not.
what log likelihood value says ? whether it should be low or high ?
because i used this model to predict the 4 category of response variable by
choosing those datapoint which were used to fit the model.
i get 72% of training data ( those which were used to fit model) correctly
predicted.
please help
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
