Error in UseMethod("compute"): no applicable method for 'compute' applied to an object of class "nn"

Question

Getting the error while running compute() from the neuralnet package in R.

Is it happening because of Data Size? I can't figure out the exact problem.

 df2 <- read.csv("data.csv")
 train_df <- df2[1:3200,]
 test_df <- df2[3201:4004,]

 n <- names(train_df)
 f <- as.formula(paste("tenure ~", paste(n[!n %in% "tenure"], collapse = 
                  "+")))

 model2 <- neuralnet(f,train_df, hidden=3, threshold=0.01, linear.output=T)

 summary(model2)

 #Output
                   Length  Class      Mode    
 call                      6 -none-     call    
 response               3200 -none-     numeric 
 covariate           4118400 -none-     numeric 
 model.list                2 -none-     list    
 err.fct                   1 -none-     function
 act.fct                   1 -none-     function
 linear.output             1 -none-     logical 
 data                   1288 data.frame list    
 net.result                1 -none-     list    
 weights                   1 -none-     list    
 startweights              1 -none-     list    
 generalized.weights       1 -none-     list    
  result.matrix          3871 -none-     numeric 


 results <- compute(model2, test_df)

 #Error
 Error in UseMethod("compute"): no applicable method for 'compute' applied 
 to an object of class "nn"
 Traceback:

 1. compute(model2, test_df)

P.S: Data Columns are numeric.

Answer 1

Answer

You have multiple packages loaded that contain the compute function, and consequently you are using the wrong one. Force using compute from the neuralnet package:

results <- neuralnet::compute(model2, test_df)

Reasoning

The error says it used the line UseMethod("compute"). This line of code is not present in neuralnet::compute. Therefore, you seem to be using compute from a different package. This can happen when you load the neuralnet package followed by another package that contains a compute function (like the dplyr package). You can avoid this by using ::: neuralnet::compute.

Additional information

With find, you can find all the namespaces in which your function is defined, as well as the order in which R will look through the namespaces:

find("compute")
# [1] "package:neuralnet" "package:dplyr"