I am using count regression models for analysis of count data. finally I selected hurdle Poisson regression model by using model selection criteria. by default, the hurdle model result has two parts: count part and zero inflated(logistic part). So, which part can I use in interpreting the result? and what is the difference between the two?
I apologize but can you describe or give references to what you did. Thanks, David Booth
dear David booth, thanks for your concern. my research tittle is determining factors affecting desired number of antenatal care visits. by nature the response variable is discrete and there is excess zeros .By using model selection criteria like voung, LRT, AIC and BIC , I preferred hurdle Poisson model and it has count and zero inflated part. BUT I am confused in differentiating the two parts(count and zero inflated part).
OK I am attaching a paper of ours. Please read the first two or so sections and tell me:
1. Are you building an explanatory model or a predictive model? and 2. How did you decide on the model you are asking about? After you do the that I think we will be able to help you. Best wishes, David Booth
- Badges
- Science topic
- Similar topics
- Papers