https://justinstone.online/Justin StoneI build practical systems across software, automation, local AI tooling, data workflows, and technical problem-solving. 2026-06-25T04:08:27-05:00 Justin Stone https://justinstone.online/ Jekyll © 2026 Justin Stone /assets/img/favicons/favicon.ico /assets/img/favicons/favicon-96x96.png Quality, Then Area, Then Everything Else: Ames Housing with Stacked Ensembles2026-04-24T00:00:05-05:00 2026-06-24T23:24:28-05:00 https://justinstone.online/posts/house-prices-ames-stacked-ensembles/ Justin Stone A stacked Ridge + XGBoost + LightGBM ensemble on Ames houses reaches 0.126 RMSLE, beating a Ridge baseline by about 8 percent. The stack weights are unbalanced, and that's fine. The Default Threshold Is the Bug: Credit Card Fraud on the ULB Benchmark2026-04-24T00:00:00-05:00 2026-04-24T02:34:45-05:00 https://justinstone.online/posts/credit-card-fraud-default-threshold/ Justin Stone A fraud classifier reaches 0.98 ROC-AUC on the famous ULB dataset and still ships a policy most fraud teams would reject. The 0.5 decision threshold is the bug. The California Housing Dataset Is Mostly a Geography Lesson2026-04-21T00:00:00-05:00 2026-06-24T23:24:28-05:00 https://justinstone.online/posts/california-housing-prices-dataset/ Justin Stone A practical look at the California housing prices dataset: what the rows actually mean, why the target is capped, and what this dataset is good for teaching.