Run Class Imbalance Tournament — run_imbalance

Trains LightGBM models across different class imbalance strategies (Standard, SMOTE, Adasyn, etc.) using sliding time windows. Evaluates performance using PR-AUC and calculates statistical significance. Includes common-sense hyperparameter defaults to prevent overfitting.

Usage

run_imbalance_tournament(
  tasks,
  windows,
  feature_prefix,
  bucket_name = "lake",
  inputs_prefix = "baf-fraud/05_model_input"
)

Arguments

tasks: A tibble containing recipe_name, data_folder, and scale_pos_weight.
windows: A tibble containing window_id, train_months, and test_month.
feature_prefix: Character. The upstream dependency prefix (used to force DAG execution).
bucket_name: Character. Bucket name. Default "lake".
inputs_prefix: Character. The folder containing the sampled data. Default "05_model_input".

Value

A tibble with the summarized tournament results.