Surveillance data are ubiquitous in Latin American countries, but most times of poor quality. I am currently using a mixed-effects model, but should I use Poisson or Linear? I know there are a lot of alternatives for analysis of time series data but these two have the strength that they are the ones that are easier to communicate to the decision makers. Thanks in advance for your response.
Surveillance data in a particular context are likely to be rare because of their confidential nature. Rare events are often modelled by Poisson distributions. Therefore apart from the ease of interpretation they offer, theoretically Poisson regressions are most adequate for such data.
My Dear Antanis... firstly I hope all success to you in your work.
You can use some certieria to select the best regression fit by AIC, MSE, P_value ,..
With Best Wishes
- Badges
- Science topic
- Similar topics
- Computer Science
- Program