find_model_through_bayes: Hyper parameter search using bayesian optimisation
In jeroenvdhoven/datapiper: datapiper

Description Usage Arguments Details Value

Hyper parameter search using bayesian optimisation

find_model_through_bayes(train, test, response,
  preprocess_pipes = list(function(train, test) return(list(train =
  train, test = train, .predict = function(data) return(data)))), models,
  metrics, target_metric, higher_is_better, N_init = 20,
  N_experiment = 40, sigma_noise = 1e-08, prepend_data_checker = T,
  on_missing_column = c("error", "add")[1],
  on_extra_column = c("remove", "error")[1],
  on_type_error = c("ignore", "error")[1], seed = 1, verbose = T,
  save_model = F)

`train`	The training dataset
`test`	The testing dataset
`response`	The response column as a string
`preprocess_pipes`	List of preprocessing pipelines generated using `pipeline`.
`models`	A list of models. Each model should be a list, containing at least a training function `.train` and a `.predict` function, plus named vectors of parameters to explore. The `.train` function has to take a `data` argument that stores the training data and a `...` argument for the parameters. The `.predict` function needs to take two arguments, where the first is the model and the second the new dataset. If a parameter only takes a single value, you can use a vector to store options. Otherwise use a list. You can use `model_trainer` as a wrapper for this list. It will also test your inputs.
`metrics`	A list of metrics (functions) that need to be calculated on the train and test response and predictions. Must be named.
`target_metric`	The name of the metric to optimise. Optimisation will be done on the testset performance of this metric.
`higher_is_better`	A flag indicating if a high value of `target_metric` indicates a good result.
`N_init`	Number of iterations to initialise the bayesian optimisation with.
`N_experiment`	Number of experimentations done with the bayesian optimisation. A numeric vector with as many entries as `x`. A numeric matrix with as many columns as entries in `x`.
`sigma_noise`	An estimate of the inherent noise in sampling from. If this is set below 1e-8, we will not reconsider previously tried configurations.
`prepend_data_checker`	Flag indicating if `pipe_check` should be prepended before all pipelines.
`on_missing_column`	See `pipe_check` for details.
`on_extra_column`	See `pipe_check` for details.
`on_type_error`	See `pipe_check` for details.
`seed`	Random seed to set each time before a model is trained. Set this to 0 to ignore setting seeds.
`verbose`	Should intermediate updates be printed.
`save_model`	Flag indicating if the generated models should be saved. Defaults to False.