Computational Model for Homeowner Choice Optimization

Computational Model for Homeowner Choice Optimization Authors: Tanjeel Ahmed Bin Zaman; Tabassum Haque; Himadri Sen Gupta; Mojtaba Harati; Niamat Ullah Ibne Hossain; John W. van de Lindt This deposit provides materials to reproduce the results of the manuscript “Optimal design of flood mitigation programs under behavioral uncertainty.” The paper develops a prescriptive analytics pipeline that (i) predicts homeowner mitigation uptake using a CatBoost classifier with post-hoc Dirichlet calibration, and (ii) embeds calibrated probabilities in a constrained optimization framework to select cost-effective and (optionally) equity-aware flood mitigation program menus. Provenance and scope Source data: The survey data used in this study are publicly available via DesignSafe-CI (DOI: 10.17603/ds2-ebpq-6s93). Study context: Stated-preference discrete choice experiment with 1,366 U.S. coastal property owners across three waves (2018–2019). Behavioral model: Gradient-boosted decision trees (CatBoost) + Dirichlet calibration to improve probability quality for downstream decision-making. Optimization: Enumeration-based selection over 30 candidate policy menus, maximizing expected mitigation uptake under budget constraints, with optional equity constraints based on income-group uptake. Key reported outcomes: The calibrated model reaches 90.9% accuracy and 0.240 log loss, and the optimized uptake–budget relationship exhibits a stepwise frontier with diminishing returns at higher budgets. Contents Manuscript.pdf-final paper PDF (this manuscript). README.md-how to run the replication (expected runtime, hardware notes, random seeds). environment.yml (or requirements.txt)-pinned Python environment for reproducibility. notebooks/ 01_preprocess.ipynb-feature engineering + train/test splits (grouped CV). 02_model_train_calibrate.ipynb-CatBoost training + Dirichlet calibration diagnostics (log loss, etc.). 03_menu_optimization.ipynb — menu evaluation + budget frontier + equity constrained runs. data/ raw/ — (optional) pointers only (or a text file) directing users to download raw data from DesignSafe DOI. processed/ — analysis-ready tables derived from the raw download (if you’re allowed to redistribute). outputs/ exported tables/figures (frontier tables, calibration plots, menu selections, etc.) Data availability The underlying survey dataset is available at DesignSafe-CI: DOI 10.17603/ds2-ebpq-6s93.