R/ml_classification_decision_tree_classifier.R

In sparklyr: R Interface to Apache Spark

#' @include ml_clustering.R
NULL

#' @rdname ml_decision_tree
#' @template roxlate-ml-probabilistic-classifier-params
#' @export
ml_decision_tree_classifier <- function(
  x,
  formula = NULL,
  max_depth = 5,
  max_bins = 32,
  min_instances_per_node = 1,
  min_info_gain = 0,
  impurity = "gini",
  seed = NULL,
  thresholds = NULL,
  cache_node_ids = FALSE,
  checkpoint_interval = 10,
  max_memory_in_mb = 256,
  features_col = "features",
  label_col = "label",
  prediction_col = "prediction",
  probability_col = "probability",
  raw_prediction_col = "rawPrediction",
  uid = random_string("decision_tree_classifier_"),
  ...
) {
  check_dots_used()
  UseMethod("ml_decision_tree_classifier")
}

#' @export
ml_decision_tree_classifier.spark_connection <- function(
  x,
  formula = NULL,
  max_depth = 5,
  max_bins = 32,
  min_instances_per_node = 1,
  min_info_gain = 0,
  impurity = "gini",
  seed = NULL,
  thresholds = NULL,
  cache_node_ids = FALSE,
  checkpoint_interval = 10,
  max_memory_in_mb = 256,
  features_col = "features",
  label_col = "label",
  prediction_col = "prediction",
  probability_col = "probability",
  raw_prediction_col = "rawPrediction",
  uid = random_string("decision_tree_classifier_"),
  ...
) {
  .args <- list(
    max_depth = max_depth,
    max_bins = max_bins,
    min_instances_per_node = min_instances_per_node,
    min_info_gain = min_info_gain,
    impurity = impurity,
    seed = seed,
    thresholds = thresholds,
    cache_node_ids = cache_node_ids,
    checkpoint_interval = checkpoint_interval,
    max_memory_in_mb = max_memory_in_mb,
    features_col = features_col,
    label_col = label_col,
    prediction_col = prediction_col,
    probability_col = probability_col,
    raw_prediction_col = raw_prediction_col
  ) %>%
    c(rlang::dots_list(...)) %>%
    validator_ml_decision_tree_classifier()

  jobj <- spark_pipeline_stage(
    x,
    "org.apache.spark.ml.classification.DecisionTreeClassifier",
    uid,
    features_col = .args[["features_col"]],
    label_col = .args[["label_col"]],
    prediction_col = .args[["prediction_col"]],
    probability_col = .args[["probability_col"]],
    raw_prediction_col = .args[["raw_prediction_col"]]
  ) %>%
    invoke("setCheckpointInterval", .args[["checkpoint_interval"]]) %>%
    invoke("setImpurity", .args[["impurity"]]) %>%
    invoke("setMaxBins", .args[["max_bins"]]) %>%
    invoke("setMaxDepth", .args[["max_depth"]]) %>%
    invoke("setMinInfoGain", .args[["min_info_gain"]]) %>%
    invoke("setMinInstancesPerNode", .args[["min_instances_per_node"]]) %>%
    invoke("setCacheNodeIds", .args[["cache_node_ids"]]) %>%
    invoke("setMaxMemoryInMB", .args[["max_memory_in_mb"]]) %>%
    jobj_set_param("setThresholds", .args[["thresholds"]]) %>%
    jobj_set_param("setSeed", .args[["seed"]])

  new_ml_decision_tree_classifier(jobj)
}

#' @export
ml_decision_tree_classifier.ml_pipeline <- function(
  x,
  formula = NULL,
  max_depth = 5,
  max_bins = 32,
  min_instances_per_node = 1,
  min_info_gain = 0,
  impurity = "gini",
  seed = NULL,
  thresholds = NULL,
  cache_node_ids = FALSE,
  checkpoint_interval = 10,
  max_memory_in_mb = 256,
  features_col = "features",
  label_col = "label",
  prediction_col = "prediction",
  probability_col = "probability",
  raw_prediction_col = "rawPrediction",
  uid = random_string("decision_tree_classifier_"),
  ...
) {
  stage <- ml_decision_tree_classifier.spark_connection(
    x = spark_connection(x),
    formula = formula,
    max_depth = max_depth,
    max_bins = max_bins,
    min_instances_per_node = min_instances_per_node,
    min_info_gain = min_info_gain,
    impurity = impurity,
    seed = seed,
    thresholds = thresholds,
    cache_node_ids = cache_node_ids,
    checkpoint_interval = checkpoint_interval,
    max_memory_in_mb = max_memory_in_mb,
    features_col = features_col,
    label_col = label_col,
    prediction_col = prediction_col,
    probability_col = probability_col,
    raw_prediction_col = raw_prediction_col,
    uid = uid,
    ...
  )

  ml_add_stage(x, stage)
}

#' @export
ml_decision_tree_classifier.tbl_spark <- function(
  x,
  formula = NULL,
  max_depth = 5,
  max_bins = 32,
  min_instances_per_node = 1,
  min_info_gain = 0,
  impurity = "gini",
  seed = NULL,
  thresholds = NULL,
  cache_node_ids = FALSE,
  checkpoint_interval = 10,
  max_memory_in_mb = 256,
  features_col = "features",
  label_col = "label",
  prediction_col = "prediction",
  probability_col = "probability",
  raw_prediction_col = "rawPrediction",
  uid = random_string("decision_tree_classifier_"),
  response = NULL,
  features = NULL,
  predicted_label_col = "predicted_label",
  ...
) {
  formula <- ml_standardize_formula(formula, response, features)

  stage <- ml_decision_tree_classifier.spark_connection(
    x = spark_connection(x),
    formula = formula,
    max_depth = max_depth,
    max_bins = max_bins,
    min_instances_per_node = min_instances_per_node,
    min_info_gain = min_info_gain,
    impurity = impurity,
    seed = seed,
    thresholds = thresholds,
    cache_node_ids = cache_node_ids,
    checkpoint_interval = checkpoint_interval,
    max_memory_in_mb = max_memory_in_mb,
    features_col = features_col,
    label_col = label_col,
    prediction_col = prediction_col,
    probability_col = probability_col,
    raw_prediction_col = raw_prediction_col,
    uid = uid,
    ...
  )

  if (is.null(formula)) {
    stage %>%
      ml_fit(x)
  } else {
    ml_construct_model_supervised(
      new_ml_model_decision_tree_classification,
      predictor = stage,
      formula = formula,
      dataset = x,
      features_col = features_col,
      label_col = label_col,
      predicted_label_col = predicted_label_col
    )
  }
}

validator_ml_decision_tree_classifier <- function(.args) {
  .args <- ml_validate_decision_tree_args(.args)
  .args[["thresholds"]] <- cast_double_list(
    .args[["thresholds"]],
    allow_null = TRUE
  )
  .args[["impurity"]] <- cast_choice(.args[["impurity"]], c("gini", "entropy"))
  .args
}

new_ml_decision_tree_classifier <- function(jobj) {
  new_ml_probabilistic_classifier(jobj, class = "ml_decision_tree_classifier")
}

new_ml_decision_tree_classification_model <- function(jobj) {
  new_ml_probabilistic_classification_model(
    jobj,
    # `depth` and `featureImportances` are lazy vals in Spark.
    depth = function() invoke(jobj, "depth"),
    feature_importances = possibly_null(
      ~ read_spark_vector(jobj, "featureImportances")
    ),
    # `numNodes` is a def in Spark.
    num_nodes = function() invoke(jobj, "numNodes"),
    class = "ml_decision_tree_classification_model"
  )
}