CODEWITHSUNDEEP
rrasterrandom-forestpredictionr-raster
loki
loki

Reputation: 10350

Changing names of variables in randomForest object

I want to predict a randomForest object to a huge RasterStack.
The randomForestobject was trained with a dataframe with vairables named like "05_absor_1", "05_absor_2" ...
The RasterStack has 189 layers, which are names like "X05_absor_1", "X05_absor_2" ...
I tried to change the names of the layers of the RasterStack, but, as it occurs, they cannot be named beginning with a number.

Is there any possibility to adapt these two data sets to each other by renaming the variable names of the RandomForest object?

Upvotes: 2

Views: 1297

Answers (1)

Muesgen
Muesgen

Reputation: 162

Better late then never

You can't rename the variables in the randomForest object, but you can extract the Raster stack. Then you can change the colnames, predict the model on the exracted dataframe. At the end you can overwrite the raster cell values of one rasterlayer in the stack with the values of the predicted data frame. This will present your prediciton on the raster stack.

Look at this post, facing a similar issue with spaces in the model variable names:

https://stackoverflow.com/a/66301739/13440906

Upvotes: 1

Related Questions