I'm modeling claim count using glm. My data contains both continuous and categorical data. Data is aggregated to 6000 risk profiles(referenced here by ID). Exposure for a policyholder for 1 year

enter image description here

Here a summary of all columns: enter image description here

In fact, As far as I know, I must probably convert my continuous variables to categorical, is that all I have to do? I'll be modeling my glm using first Poisson and N.B type 2, I also think I might use exposure as an offset? Also, should ID and Year, stay as an explanatory variable?

  • $\begingroup$ you need to look at the data before deciding whether to convert it to categorical. It's not a must. also some of your columns, e.g Rent are character.. hope that is intended. ID is a unique entry, you cannot use it $\endgroup$
    – StupidWolf
    May 29, 2020 at 9:21
  • $\begingroup$ thanks for your comment, but if they are caracters, must it be converted to numbers or ligical when its 2 choices? $\endgroup$
    – roger
    May 29, 2020 at 10:43

1 Answer 1


From my understanding, transfomring categorical to dummy variable may not help you in the end.


This article shows specific steps about poisson regression modelling process in considering overdispersion and engineering variables.

Using offset may help to address over-dispersion, the article is worth a good read, after using offset, the model appearted to be good fit.

Regards, Dan


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.