Knn and nn models not saving as variables

#1

I'm building several models. My large Knn, ada and nn models are not saving as variables after they are trained. My glm, and dtree models do save. When I open R Studio again, I have to rerun the models which takes hours. Here is an example knn model.

Cross validation 10-fold CV

cross_validation <- trainControl(method = "repeatedcv",
number = 10,
repeats = 3)

Hyper-parameter tuning

Param_Grid <- expand.grid(k = 1:10)

fit the model to training data

Knn_prev <- train(HN_data_x_train_prev_norm,
HN_data_y_train_1,
method = "knn",
tuneGrid = Param_Grid,
trControl = cross_validation)

#2

What do you mean...
That Knn_prev doesn't have content when the train completes? Does train error?

Or do you mean that you dont know how to save it to disk?
You can try base::save or base::saveRDS

#4

Im surprised you shared output for your successful code but not for the code you are having trouble with...

Can you identify with me what package this train function is from?

#5

The code I'm having trouble with is in the first message. The model trains but does not save as variable "Knn_prev".

#6

What is the result of

getAnywhere(train)
#7

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.

#3

Correct. Knn_prev doesn't save any content when train completes. Train does not error. It prints the intended model. I attached my glm_fit_SMOTE train and it shows as saved data to the right.

image
image990×177 18.1 KB