I’m trying to determinate the ABG (Above ground biomass) using the random forest algorithm (and R package) on my data set. And it seems that the package can handle only few categories (when I have 53). Is there anything to do to make it work? I have read some articles about predicting ABG using RF package so I’m suprised if its not possible. I always used it with only few class so I don’t really know what to do.

My train data is field data where we have estimated the ABG by ourself.

Tune.RF<-tuneRF(train.df,train.df$Class, ntreeTry = 500, stepFactor = 2, trace = FALSE,improve = 0.5, plot = FALSE, doBest = FALSE)
Error in randomForest.default(x, y, mtry = mtryStart, ntree = ntreeTry,  : 
  Can not handle categorical predictors with more than 53 categories.

print(train.df)
         Class
1  0,000105569
2     5,18E-07
3  0,000309691
4     2,00E-06
5     1,02E-06
6   0,00020349
7  0,000128096
8  0,000438897
9  0,000236575
10 0,000158402
...

I have no clue what to do.