Can't use raster::predict function with dbart model object

Machine Learning and Modeling
,
1

Hi,

I am running two machine-learning models using caret and dbarts packages to train SVM and BART respectively. The goal is to apply these models spatially and produce a raster prediction of a continuous target variable.

The code below was used to train the BART model using parsnip:
bt_reg <-
set_mode("regression") %>%
set_engine("dbarts")

bt_reg <-
parsnip::bart() %>%
set_engine("dbarts") %>%
set_mode("regression")
set.seed(7)
bt_reg_fit <- bt_reg %>% fit(Y ~ X1+X2,
data = train_split)

I've been trying to predict a raster in R using BART and SVM models. raster::predict(rasterstack, SVM_model) work fine. But when I try to predict with BART model using raster::predict(rasterstack, BART_model), I get an error:

Error in rep_len(object$sigma, n.obs * length(object$sigma)) :
invalid 'length.out' value
In addition: Warning messages:
1: In n.obs * length(object$sigma) : NAs produced by integer overflow
2: In n.obs * length(object$sigma) : NAs produced by integer overflow

Any help would be appreciated.
Thank you

Sincerely

2

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.