This might seem a stupid question but I will ask anyway. I understand the importance of normalizing environmental data measured on different scales. Is it necessary however to normalize a categorical variable (land use data with 12 different classes)? When I plot the normalized variable the map looks the same and has retained all of its' classes. Am I doing the right thing or am I going down the wrong path regarding this.
Many thanks,
NIkhail
Nikhail
I would recommend reading the Alan Agresti book "Categorial Data Analysis". In that book, likely, you wil find the best path for your research. You can find something about this book in WEB. Best Regards
Nilson Campos
Nikhail,
I think the answer can be really short. There is no need to normalize categorical variables. You are not very explicit about the type of analysis you are doing, but typically you are dealing with the categorical variables as dummy variables in the statistical analysis.
Good luck with the analysis,
Jetse
You may have a look at my endemics and bear articles for continuous versus categorical variables in statistical spatial modeling. A case is described (soil variables) were and why fewer (lumped) categories provide better stats. Further, the papers show reducing the number of polygons/rasters within layers of categorical variables (soil; cover) by lumping/joining small polygons to neighboring ones. The reduction of the number of categories depending on spatial resolution is also argued in some other papers both for statistical and ecological reasons.
- Badges
- Science topic