fit_best.md
In tune: Tidy Tuning Tools

fit_best

Code
  fit_best(knn_pca_res, verbose = TRUE)
Output
  Using rmse as the metric, the optimal parameters were:
    neighbors: 10
    num_comp:  3

Message
  i Fitting using 161 data points...
  v Done.
Output
  == Workflow [trained] ==========================================================
  Preprocessor: Recipe
  Model: nearest_neighbor()

  -- Preprocessor ----------------------------------------------------------------
  1 Recipe Step

  * step_pca()

  -- Model -----------------------------------------------------------------------

  Call:
  kknn::train.kknn(formula = ..y ~ ., data = data, ks = min_rows(10L,     data, 5))

  Type of response variable: continuous
  minimal mean absolute error: 1.690086
  Minimal mean squared error: 4.571625
  Best kernel: optimal
  Best k: 10

Code
  tmp <- fit_best(knn_pca_res, verbose = TRUE, parameters = tibble(neighbors = 1,
    num_comp = 1))
Message
  i Fitting using 161 data points...
  v Done.

There is no `fit_best()` method for an object with class `integer`.

There is no `fit_best()` method for an object with classes `tbl_df`, `tbl`, and `data.frame`.

"WAT" was not in the metric set. Please choose from: "rmse" and "rsq".

The parameters `neighbors` and `num_comp` are still marked for tuning.

The parameter `num_comp` is still marked for tuning.

`...` must be empty.
x Problematic argument:
* chickens = 2

x The control option `save_workflow = TRUE` should be used when tuning.

fit_best() warns when metric or eval_time are specified in addition to parameters

Code
  manual_wf <- fit_best(res, metric = "rmse", parameters = tune_params)
Condition
  Warning:
  `metric` is being ignored because `parameters` has been specified.

Code
  manual_wf <- fit_best(res, metric = "rmse", eval_time = 10, parameters = tune_params)
Condition
  Warning:
  `metric` is being ignored because `parameters` has been specified.
  Warning:
  `eval_time` is being ignored because `parameters` has been specified.