In my opinion residual plots are suitable here.
ggfortify has a decent autoplot method for lm objects.
autoplot.LearnerRegrLM = function(object, ...) { # nolint
if (is.null(object$model)) {
stopf("Learner '%s' must be trained first", object$id)
}
if (is.null(object$model$model)) {
stopf("Learner '%s' must be trained with `keep_model` set to `TRUE`", object$id)
}
require_namespaces("ggfortify")
autoplot(object$model, ...)
}
In my opinion residual plots are suitable here. ggfortify has a decent autoplot method for lm objects.
usage:
An example of the plot:![lm](https://user-images.githubusercontent.com/35730264/89646294-d1e2b100-d8bb-11ea-8fb5-325cd6f5f769.png)