Automated Machine Learning (AutoML)

Code accompanying two papers:

1. Automated Machine Learning (AutoML) in Insurance, an end-to-end AutoML pipeline integrating data encoding, imputation, scaling, resampling for rare-event imbalance, and CASH-based hyperparameter optimization with stacking/bagging/boosting ensembles.
2. Starting Off on the Wrong Foot: Pitfalls in Data Preparation — the Informed Data Preparation Pipeline (IDPP) with SPlit support-point partitioning, Chatterjee correlation feature selection, and MissForest imputation, embedded in the InformedAutoML framework.

The repository implements:

• InsurAutoML and InformedAutoML pipelines
• Illustrative examples for applications (replicate of experiments in the papers)

Available at: https://github.com/PanyiDong/InsurAutoML