Bayesian optimization with transfer learning for small-sample formulation development of BCS Class II drugs: a benchmark study

This dataset contains the replication package for the study "Bayesian optimization with transfer learning for small-sample formulation development of BCS Class II drugs: a benchmark study". It includes the Python source code (tlbo_benchmark.py), dependency configuration (requirements.txt), user documentation (README.md), and a compressed archive containing all generated figures (Figs. 2–6), benchmark data tables (Tables 2, 5, and 6), and the meta-priors file (meta_priors.json). The code implements the TL-BO framework and benchmark comparisons against D-optimal DoE, standard BO, Latin hypercube sampling, and random search for griseofulvin tablet formulation.