A Julia package for empirical Bayes estimation. See the documentation for instructions on how to use it.
The package implements the empirical Bayes cross-fit method [1], which estimates effect sizes of many experiments by optimally synthesizing experimental data and rich covariate information. Furthermore, the method may leverage any black-box predictive model: [1] provides theoretical guarantees that hold for any regression method and the package here allows usage of any supervised model that has implemented the MLJ.jl interface.
[1] Ignatiadis, N., & Wager, S. (2019). Covariate-Powered Empirical Bayes Estimation. To appear in Advances in Neural Information Processing Systems 32 (NeurIPS 2019). arXiv:1906.01611.