kiddo
06-11-2010, 09:48 AM
hello there!
im currently doing a paper which involves count data. i used both Poisson Regression and Negative Binomial Regression. i want to know which of the 2 models provided a better fit by comparing them & by getting their own goodness of fit.
im planning to do something like this
count observed POisson Negative Binomial
0 100 - -
1 12 - -
2 13 - -
3 26 - -
and then i will get their goodness of fit.
with that, i would like to ask how can i get the expected frequencies for count=0,1,2,3 using Poisson & Negative Binomial.
the data given to me was raw data with more than 20 explanatory (independent) variables and not the contingency table.
additional info: i already acquired the estimates for the model but i dont know how i can use that to get the expected frequency for count=0,1,2,3.
hope you can help me. thank you
im currently doing a paper which involves count data. i used both Poisson Regression and Negative Binomial Regression. i want to know which of the 2 models provided a better fit by comparing them & by getting their own goodness of fit.
im planning to do something like this
count observed POisson Negative Binomial
0 100 - -
1 12 - -
2 13 - -
3 26 - -
and then i will get their goodness of fit.
with that, i would like to ask how can i get the expected frequencies for count=0,1,2,3 using Poisson & Negative Binomial.
the data given to me was raw data with more than 20 explanatory (independent) variables and not the contingency table.
additional info: i already acquired the estimates for the model but i dont know how i can use that to get the expected frequency for count=0,1,2,3.
hope you can help me. thank you