xgboost: Regression

# nolint start
library(mlexperiments)
library(mllrnrs)

See https://github.com/kapsner/mllrnrs/blob/main/R/learner_xgboost.R for implementation details.

Preprocessing

Import and Prepare Data

library(mlbench)
data("BostonHousing")
dataset <- BostonHousing |>
  data.table::as.data.table() |>
  na.omit()

feature_cols <- colnames(dataset)[1:13]
target_col <- "medv"

General Configurations

seed <- 123
if (isTRUE(as.logical(Sys.getenv("_R_CHECK_LIMIT_CORES_")))) {
  # on cran
  ncores <- 2L
} else {
  ncores <- ifelse(
    test = parallel::detectCores() > 4,
    yes = 4L,
    no = ifelse(
      test = parallel::detectCores() < 2L,
      yes = 1L,
      no = parallel::detectCores()
    )
  )
}
options("mlexperiments.bayesian.max_init" = 4L)
options("mlexperiments.optim.xgb.nrounds" = 20L)
options("mlexperiments.optim.xgb.early_stopping_rounds" = 5L)

Generate Training- and Test Data

data_split <- splitTools::partition(
  y = dataset[, get(target_col)],
  p = c(train = 0.7, test = 0.3),
  type = "stratified",
  seed = seed
)

train_x <- model.matrix(
  ~ -1 + .,
  dataset[data_split$train, .SD, .SDcols = feature_cols]
)
train_y <- log(dataset[data_split$train, get(target_col)])


test_x <- model.matrix(
  ~ -1 + .,
  dataset[data_split$test, .SD, .SDcols = feature_cols]
)
test_y <- log(dataset[data_split$test, get(target_col)])

Generate Training Data Folds

fold_list <- splitTools::create_folds(
  y = train_y,
  k = 3,
  type = "stratified",
  seed = seed
)

Experiments

Prepare Experiments

# required learner arguments, not optimized
learner_args <- list(
  objective = "reg:squarederror",
  eval_metric = "rmse"
)

# set arguments for predict function and performance metric,
# required for mlexperiments::MLCrossValidation and
# mlexperiments::MLNestedCV
predict_args <- NULL
performance_metric <- metric("rmsle")
performance_metric_args <- NULL
return_models <- FALSE

# required for grid search and initialization of bayesian optimization
parameter_grid <- expand.grid(
  subsample = seq(0.6, 1, .2),
  colsample_bytree = seq(0.6, 1, .2),
  min_child_weight = seq(1, 5, 4),
  learning_rate = seq(0.1, 0.2, 0.1),
  max_depth = seq(1, 5, 4)
)
# reduce to a maximum of 10 rows
if (nrow(parameter_grid) > 10) {
  set.seed(123)
  sample_rows <- sample(seq_len(nrow(parameter_grid)), 10, FALSE)
  parameter_grid <- kdry::mlh_subset(parameter_grid, sample_rows)
}

# required for bayesian optimization
parameter_bounds <- list(
  subsample = c(0.2, 1),
  colsample_bytree = c(0.2, 1),
  min_child_weight = c(1L, 10L),
  learning_rate = c(0.1, 0.2),
  max_depth =  c(1L, 10L)
)
optim_args <- list(
  n_iter = ncores,
  kappa = 3.5,
  acq = "ucb"
)

Hyperparameter Tuning

tuner <- mlexperiments::MLTuneParameters$new(
  learner = mllrnrs::LearnerXgboost$new(
    metric_optimization_higher_better = FALSE
  ),
  strategy = "grid",
  ncores = ncores,
  seed = seed
)

tuner$parameter_grid <- parameter_grid
tuner$learner_args <- learner_args
tuner$split_type <- "stratified"

tuner$set_data(
  x = train_x,
  y = train_y
)

tuner_results_grid <- tuner$execute(k = 3)
#>
#> Parameter settings [=====================================>----------------------------------------------------------] 4/10 ( 40%)
#> Parameter settings [===============================================>------------------------------------------------] 5/10 ( 50%)
#> Parameter settings [=========================================================>--------------------------------------] 6/10 ( 60%)
#> Parameter settings [==================================================================>-----------------------------] 7/10 ( 70%)
#> Parameter settings [============================================================================>-------------------] 8/10 ( 80%)
#> Parameter settings [=====================================================================================>----------] 9/10 ( 90%)
#> Parameter settings [===============================================================================================] 10/10 (100%)

head(tuner_results_grid)
#>    setting_id metric_optim_mean nrounds subsample colsample_bytree min_child_weight learning_rate max_depth        objective
#>         <int>             <num>   <int>     <num>            <num>            <num>         <num>     <num>           <char>
#> 1:          1         0.1872223      62       0.6              0.8                5           0.2         1 reg:squarederror
#> 2:          2         0.1688788      99       1.0              0.8                5           0.1         5 reg:squarederror
#> 3:          3         0.1916676      97       0.8              0.8                5           0.1         1 reg:squarederror
#> 4:          4         0.1662343      55       0.6              0.8                5           0.2         5 reg:squarederror
#> 5:          5         0.1635528     100       1.0              0.8                1           0.1         5 reg:squarederror
#> 6:          6         0.1641982     100       0.8              0.8                5           0.1         5 reg:squarederror
#>    eval_metric
#>         <char>
#> 1:        rmse
#> 2:        rmse
#> 3:        rmse
#> 4:        rmse
#> 5:        rmse
#> 6:        rmse

Bayesian Optimization

tuner <- mlexperiments::MLTuneParameters$new(
  learner = mllrnrs::LearnerXgboost$new(
    metric_optimization_higher_better = FALSE
  ),
  strategy = "bayesian",
  ncores = ncores,
  seed = seed
)

tuner$parameter_grid <- parameter_grid
tuner$parameter_bounds <- parameter_bounds

tuner$learner_args <- learner_args
tuner$optim_args <- optim_args

tuner$split_type <- "stratified"

tuner$set_data(
  x = train_x,
  y = train_y
)

tuner_results_bayesian <- tuner$execute(k = 3)
#>
#> Registering parallel backend using 4 cores.

head(tuner_results_bayesian)
#>    Epoch setting_id subsample colsample_bytree min_child_weight learning_rate max_depth gpUtility acqOptimum inBounds Elapsed
#>    <num>      <int>     <num>            <num>            <num>         <num>     <num>     <num>     <lgcl>   <lgcl>   <num>
#> 1:     0          1       0.6              0.8                5           0.2         1        NA      FALSE     TRUE   0.915
#> 2:     0          2       1.0              0.8                5           0.1         5        NA      FALSE     TRUE   1.028
#> 3:     0          3       0.8              0.8                5           0.1         1        NA      FALSE     TRUE   0.935
#> 4:     0          4       0.6              0.8                5           0.2         5        NA      FALSE     TRUE   1.013
#> 5:     0          5       1.0              0.8                1           0.1         5        NA      FALSE     TRUE   0.248
#> 6:     0          6       0.8              0.8                5           0.1         5        NA      FALSE     TRUE   0.218
#>         Score metric_optim_mean nrounds errorMessage        objective eval_metric
#>         <num>             <num>   <int>       <lgcl>           <char>      <char>
#> 1: -0.1872223         0.1872223      62           NA reg:squarederror        rmse
#> 2: -0.1688788         0.1688788      99           NA reg:squarederror        rmse
#> 3: -0.1916676         0.1916676      97           NA reg:squarederror        rmse
#> 4: -0.1662343         0.1662343      55           NA reg:squarederror        rmse
#> 5: -0.1635528         0.1635528     100           NA reg:squarederror        rmse
#> 6: -0.1641982         0.1641982     100           NA reg:squarederror        rmse

k-Fold Cross Validation

validator <- mlexperiments::MLCrossValidation$new(
  learner = mllrnrs::LearnerXgboost$new(
    metric_optimization_higher_better = FALSE
  ),
  fold_list = fold_list,
  ncores = ncores,
  seed = seed
)

validator$learner_args <- tuner$results$best.setting[-1]

validator$predict_args <- predict_args
validator$performance_metric <- performance_metric
validator$performance_metric_args <- performance_metric_args
validator$return_models <- return_models

validator$set_data(
  x = train_x,
  y = train_y
)

validator_results <- validator$execute()
#>
#> CV fold: Fold1
#>
#> CV fold: Fold2
#>
#> CV fold: Fold3

head(validator_results)
#>      fold performance subsample colsample_bytree min_child_weight learning_rate max_depth nrounds        objective eval_metric
#>    <char>       <num>     <num>            <num>            <num>         <num>     <num>   <int>           <char>      <char>
#> 1:  Fold1  0.03396237       0.6                1                1           0.1         5      99 reg:squarederror        rmse
#> 2:  Fold2  0.05035231       0.6                1                1           0.1         5      99 reg:squarederror        rmse
#> 3:  Fold3  0.03987737       0.6                1                1           0.1         5      99 reg:squarederror        rmse

Nested Cross Validation

validator <- mlexperiments::MLNestedCV$new(
  learner = mllrnrs::LearnerXgboost$new(
    metric_optimization_higher_better = FALSE
  ),
  strategy = "grid",
  fold_list = fold_list,
  k_tuning = 3L,
  ncores = ncores,
  seed = seed
)

validator$parameter_grid <- parameter_grid
validator$learner_args <- learner_args
validator$split_type <- "stratified"

validator$predict_args <- predict_args
validator$performance_metric <- performance_metric
validator$performance_metric_args <- performance_metric_args
validator$return_models <- return_models

validator$set_data(
  x = train_x,
  y = train_y
)

validator_results <- validator$execute()
#>
#> CV fold: Fold1
#>
#> Parameter settings [=====================================>----------------------------------------------------------] 4/10 ( 40%)
#> Parameter settings [===============================================>------------------------------------------------] 5/10 ( 50%)
#> Parameter settings [=========================================================>--------------------------------------] 6/10 ( 60%)
#> Parameter settings [==================================================================>-----------------------------] 7/10 ( 70%)
#> Parameter settings [============================================================================>-------------------] 8/10 ( 80%)
#> Parameter settings [=====================================================================================>----------] 9/10 ( 90%)
#> Parameter settings [===============================================================================================] 10/10 (100%)
#> CV fold: Fold2
#> CV progress [====================================================================>-----------------------------------] 2/3 ( 67%)
#>
#> Parameter settings [============================>-------------------------------------------------------------------] 3/10 ( 30%)
#> Parameter settings [=====================================>----------------------------------------------------------] 4/10 ( 40%)
#> Parameter settings [===============================================>------------------------------------------------] 5/10 ( 50%)
#> Parameter settings [=========================================================>--------------------------------------] 6/10 ( 60%)
#> Parameter settings [==================================================================>-----------------------------] 7/10 ( 70%)
#> Parameter settings [============================================================================>-------------------] 8/10 ( 80%)
#> Parameter settings [=====================================================================================>----------] 9/10 ( 90%)
#> Parameter settings [===============================================================================================] 10/10 (100%)
#> CV fold: Fold3
#> CV progress [========================================================================================================] 3/3 (100%)
#>
#> Parameter settings [===============================================>------------------------------------------------] 5/10 ( 50%)
#> Parameter settings [=========================================================>--------------------------------------] 6/10 ( 60%)
#> Parameter settings [==================================================================>-----------------------------] 7/10 ( 70%)
#> Parameter settings [============================================================================>-------------------] 8/10 ( 80%)
#> Parameter settings [=====================================================================================>----------] 9/10 ( 90%)
#> Parameter settings [===============================================================================================] 10/10 (100%)

head(validator_results)
#>      fold performance nrounds subsample colsample_bytree min_child_weight learning_rate max_depth        objective eval_metric
#>    <char>       <num>   <int>     <num>            <num>            <num>         <num>     <num>           <char>      <char>
#> 1:  Fold1  0.03942935      53       0.8              0.8                5           0.1         5 reg:squarederror        rmse
#> 2:  Fold2  0.05037283     100       0.6              1.0                1           0.1         5 reg:squarederror        rmse
#> 3:  Fold3  0.04125686      35       0.6              1.0                5           0.2         5 reg:squarederror        rmse

Inner Bayesian Optimization

validator <- mlexperiments::MLNestedCV$new(
  learner = mllrnrs::LearnerXgboost$new(
    metric_optimization_higher_better = FALSE
  ),
  strategy = "bayesian",
  fold_list = fold_list,
  k_tuning = 3L,
  ncores = ncores,
  seed = seed
)

validator$parameter_grid <- parameter_grid
validator$learner_args <- learner_args
validator$split_type <- "stratified"


validator$parameter_bounds <- parameter_bounds
validator$optim_args <- optim_args

validator$predict_args <- predict_args
validator$performance_metric <- performance_metric
validator$performance_metric_args <- performance_metric_args
validator$return_models <- TRUE

validator$set_data(
  x = train_x,
  y = train_y
)

validator_results <- validator$execute()
#>
#> CV fold: Fold1
#>
#> Registering parallel backend using 4 cores.
#>
#> CV fold: Fold2
#> CV progress [====================================================================>-----------------------------------] 2/3 ( 67%)
#>
#> Registering parallel backend using 4 cores.
#>
#> CV fold: Fold3
#> CV progress [========================================================================================================] 3/3 (100%)
#>
#> Registering parallel backend using 4 cores.

head(validator_results)
#>      fold performance subsample colsample_bytree min_child_weight learning_rate max_depth nrounds        objective eval_metric
#>    <char>       <num>     <num>            <num>            <num>         <num>     <num>   <int>           <char>      <char>
#> 1:  Fold1  0.04712865 0.5029800        0.4977050                6     0.1195995         2      53 reg:squarederror        rmse
#> 2:  Fold2  0.05054316 0.4489853        0.7725962                2     0.2000000         5      65 reg:squarederror        rmse
#> 3:  Fold3  0.04027241 0.7465061        0.8234365                1     0.2000000         5      29 reg:squarederror        rmse

Holdout Test Dataset Performance

Predict Outcome in Holdout Test Dataset

preds_xgboost <- mlexperiments::predictions(
  object = validator,
  newdata = test_x
)

Evaluate Performance on Holdout Test Dataset

perf_xgboost <- mlexperiments::performance(
  object = validator,
  prediction_results = preds_xgboost,
  y_ground_truth = test_y,
  type = "regression"
)
perf_xgboost
#>     model performance      SSE        MSE      RMSE       MEDSE      SAE       MAE      MEDAE       RSQ    EXPVAR      RRSE
#>    <char>       <num>    <num>      <num>     <num>       <num>    <num>     <num>      <num>     <num>     <num>     <num>
#> 1:  Fold1  0.04135927 4.102050 0.02646484 0.1626802 0.006572713 17.66284 0.1139538 0.08107227 0.8278688 0.8280100 0.4148870
#> 2:  Fold2  0.04757084 4.799370 0.03096368 0.1759650 0.006181385 19.54728 0.1261115 0.07862178 0.7986077 0.9943971 0.4487675
#> 3:  Fold3  0.03918577 3.567998 0.02301934 0.1517213 0.006275152 17.01818 0.1097947 0.07921586 0.8502788 0.8048249 0.3869382
#>          RAE       MAPE KendallTau SpearmanRho
#>        <num>      <num>      <num>       <num>
#> 1: 0.3847300 0.03826134  0.7533943   0.9036017
#> 2: 0.4257766 0.04426443  0.7628554   0.9148171
#> 3: 0.3706880 0.03763162  0.7815047   0.9278985