[R] How to find absolute residuals

[Ivan Krylov]

On Wed, 18 Dec 2019 10:06:56 +0100
javed khan <javedbtk111 using gmail.com> wrote:

> grid_search <- train(log10(Results) ~ ., data = tr,
>                      method = "svmRadial",
>                      ## Will create 48 parameter combinations
>                      ##  tuneLength = 15,
>                      metric = "MAE",
>                      preProc = c("center", "scale", "zv"),
>                      trControl = ctrl2)

There is no "train" function in R core, so I'm going to invoke my
psychic debugging powers and determine that you are using the "caret"
package.

> I need the absolute residuals (actual-predicted values) in a table or
> another sequential format so that I may perform statistical analysis
> on it later.

Caret seems to provide a residuals() method for objects of class "train"
returned from the train() function. If all else fails, you could always
call predict() and compute the residuals yourself.

-- 
Best regards,
Ivan