R/catboost_fit.R

In statist-bhfz/grideR: Easy model tuning with data.table

#' Wrapper for catboost.train.
#'
#' Fit and evaluate catboost model with data.table as input data.
#' Model are trained (including all preprocessing steps) on train part and
#' evaluated on validation part according to \code{split} indicator variable.
#'
#' @param data data.table with all input data.
#' @param y Target variable name (character).
#' @param split Indicator variable with 1 corresponds to observations in validation dataset.
#' @param preproc_fun Preprocessing function which takes data.table \code{data}+\code{split}
#' as input and returns processed data.table with same \code{target} and \code{split} columns.
#' @param params 1-row data.table with all hyperparameters.
#' @param args NULL value for consistency with \code{xgb_fit()}.
#' @param metrics Vector of metric functions names.
#' @param return_val_preds If \code{TRUE}, predictions for validation data 
#' will be returned. 
#' @param return_model_obj If \code{TRUE}, model object will be returned.
#' @param train_on_all_data If \code{TRUE}, model will be fitted on all data
#' (without train/validation split) and model object will be returned.
#' @param ... Other parameters for \code{catboost.train()}.
#'
#' @return data.table with optimal number of iterations (if early stopping is used)
#' and all metrics calculated for validation part of the data. It also contains 
#' predictions for validation data if \code{return_val_preds = TRUE} and 
#' model object if \code{return_model_obj = TRUE}. 
#' If \code{train_on_all_data = TRUE}, only model object will be returned.
#'
#' @examples
#' # Input data
#' dt <- as.data.table(mtcars)
#' # data.table with resamples
#' splits <- resampleR::cv_base(dt, "hp")
#' # data.table with all hyperparameters
#' catboost_grid <- CJ(
#'     iterations = 1000,
#'     learning_rate = 0.05,
#'     depth = c(8, 9),
#'     loss_function = "RMSE",
#'     eval_metric = "RMSE",
#'     random_seed = 42,
#'     od_type = "Iter",
#'     # metric_period = 50,
#'     od_wait = 10,
#'     use_best_model = TRUE,
#'     logging_level = "Silent"
#' ) 
#' # Dumb preprocessing function
#' # Real function will contain imputation, feature engineering etc.
#' # with all statistics computed on train folds and applied to validation fold
#' preproc_fun_example <- function(data) return(data[])
#' catboost_fit(data = dt,
#'              target = "hp",
#'              split = splits[, split_1],
#'              preproc_fun = preproc_fun_example,
#'              params = catboost_grid[1, ],
#'              metrics = c("rmse", "mae"),
#'              return_val_preds = TRUE)
#'
#' @details
#'
#'
#' @import data.table
#' @import checkmate
#' @import catboost
#' @export
catboost_fit <- function(data = data,
                         target = target,
                         split = split,
                         preproc_fun = preproc_fun,
                         params = params,
                         args = NULL,
                         metrics = metrics,
                         return_val_preds = FALSE,
                         return_model_obj = FALSE,
                         train_on_all_data = FALSE,
                         ...) {
    
    assert_data_table(data)
    if (!train_on_all_data) assert_integerish(split, len = data[, .N])
    assert_data_table(params)
    
    data <- copy(data)[, split := split]
    data <- preproc_fun(data)
    
    if (train_on_all_data) {
        cols_to_drop <- c(target)
        dtrain <- catboost.load_pool(data[, .SD, .SDcols = -cols_to_drop],
                                     label = data[, get(target)])
        rm(data)
        args <- c(args,
                  list(params = as.list(params),
                       data = dtrain))
        model <-  catboost.train(dtrain,
                                 params = as.list(params))
        return(model)
    }
    
    val_ground_truth <- data[split == 1, get(target)]
    
    cols_to_drop <- c(target, "split")
    
    dtrain <- catboost.load_pool(data[split == 0, .SD, .SDcols = -cols_to_drop],
                                 label = data[split == 0, get(target)])
    
    dval <- catboost.load_pool(data[split == 1, .SD, .SDcols = -cols_to_drop],
                               label = data[split == 1, get(target)])
    
    rm(data)
    
    model <- catboost.train(dtrain,
                            test_pool = dval,
                            params = as.list(params))
    
    preds <- data.table(
        ground_truth = val_ground_truth,
        prediction = catboost.predict(model, dval)
    )
    
    res <- data.table(
        nrounds_best = model$tree_count
    )
    
    for (metric in metrics) {
        res[, (metric) := get(metric)(preds$ground_truth, preds$prediction)]
    }
    
    if (return_val_preds) res[, val_preds := list(list(preds[, prediction]))]
    if (return_model_obj) res[, model_obj := .(list(model))]
    
    return(res[])
}