The research behind Personalized Plan
Personalized Plan brings optimal dynamic treatment regimes, a framework from precision medicine, to high-school math course planning. Everything on this site is grounded in the work below.
Current work
The core paper. It introduces ODTRs for education and applies them to HSLS:09, finding course pathways that maximize end-of-high-school weighted math GPA under constraints that keep recommendations realistic. Both models on this site come from it.
The companion methods paper. It develops the feasibility machinery that stops the model from recommending pathways no real school would offer. The data-driven tool runs on this constrained regime.
Published related work
Earlier work by our group on optimal treatment regimes. It shapes our thinking but is not part of this site's models.
- 2026 Fair and Robust Estimation of Heterogeneous Treatment Effects for Optimal Policies in Multilevel Studies Suk, Y., Park, C., Pan, C., & Kim, K. · Multivariate Behavioral Research.
- 2023 Designing Optimal, Data-Driven Policies from Multisite Randomized Trials Suk, Y., & Park, C. · Psychometrika.
How to cite
Pan, C., Li, Y., Harris, T., & Suk, Y. (in press). Estimating optimal dynamic treatment regimes for personalized education: A tutorial and applications with machine learning. Zeitschrift für Psychologie. https://arxiv.org/abs/2610.10843
The preprint is available on arXiv. Volume, page, and DOI details will be added here once the article is published.
Research preview. The recommendations illustrate the framework and are intended to support advising conversations, not replace them.