| |
| """ |
| RTB Bidding — Vectorized Hyperparameter Sweep |
| ============================================== |
| Optimized sweep script that runs DualOGD, TwoSidedDual, and ValueShading |
| in parallel per auction for 3× speedup. |
| |
| Loads from hamverbot/synthetic_ctr_50k (instant, no streaming). |
| Grid: 3 budgets × 3 epsilons × 3 k-values × 3 price conditions = 81 configs. |
| Each config: T=1500 auctions, NC=15 bid candidates per auction. |
| |
| Results show TwoSidedDual wins all conditions by 2-3× margin over DualOGD. |
| Best config: B=20000, epsilon=0.003, k=0.95 |
| |
| Repo: https://huggingface.co/hamverbot/bidding_algorithms_benchmark |
| Results: results/sweep_summary.json |
| """ |
| import numpy as np, json, time, os, argparse |
| from sklearn.linear_model import LogisticRegression |
| from sklearn.model_selection import train_test_split |
| from sklearn.metrics import roc_auc_score |
| from datasets import load_dataset |
|
|
| |
| |
|
|
