WentaoGuo19/pswmap: pswmap 1.0.0

First release of pswmap, the R package implementing the propensity-score-weighted meta-analytic predictive (PSW-MAP) prior of Guo and Zhu for borrowing external control data into a randomized trial when baseline covariates are missing in the external data.

• pswmap() runs the full analysis in one call: source-specific propensity scores and odds weights, weighted source summaries, the robust PSW-MAP prior and the posterior of the treatment effect. borrow = FALSE analyses the trial alone.
• Missing external covariates: multiple imputation within each source (default), complete-case weighting, or the learner's native missing-data handling.
• Propensity score learners: logistic regression (default), BART, random forest and XGBoost.
• Step functions for advanced users: psw_weights(), psw_summaries(), psw_reference_scale(), psw_map_prior(), psw_posterior().
• Example data pswmap_example.

Install with remotes::install_github("WentaoGuo19/pswmap@v1.0.0").

The simulation study of the paper is in pswmap-paper.