Error in Unused Arguments

[FredOwen]

Hi all,

I have been using the example from this website without any issue at all:

https://www.r-bloggers.com/2020/05/multinomial-logistic-regression-with-r/

However there is now an error.

Here is a code snippet of the problem, once the BreastTissue data is loaded.

require(nnet)
# Training the multinomial model
multinom_model <- multinom(Class ~ ., data = tissue)

# Predicting the values for train dataset
train$ClassPredicted <- predict(multinom_model, newdata = train, "class")

Error Message below:

Error in predict(multinom_model, newdata = train, "class") : 
  unused arguments (newdata = train, "class")

It is no longer working and I have tried to delete the current version and go back a couple of versions with R studio to see if that would fix the problem, but still no luck. Very strange. Any guidance on how to solve an issue like this?

Thank you.

[FJCC]

Let's confirm that objects are what you expect. What is the result of

class(multinom_model)

I think the result should include nnet.formula.

What is the result of

train$ClassPredicted <- predict.nnet(multinom_model, newdata = train, type = "class")

[FredOwen]

I ran the following.

Result of

class(multinom_model)

is

[1] "multinom" "nnet"    

and the result of

train$ClassPredicted <- predict.nnet(multinom_model, newdata = train, type = "class")

is

Error in predict.nnet(multinom_model, newdata = train, type = "class") : 
  could not find function "predict.nnet"

[FJCC]

If you run

search()

is package::nnet in the returned values?

[FredOwen]

yes "package:nnet" is returned in the list

[FJCC]

Does

methods(predict)

show anything related to nnet?

[FredOwen]

It does. Thank you FJCC.
Endless hours later... after clearing the workspace it is now working.
Still strange as I would have expected that to work after deleting R and reinstalling etc.
I do appreciate your input.