Description
Shrinkage for Effect Estimation.
Description
Computes shrinkage estimators for regression problems. Selects penalty parameter by minimizing bias and variance in the effect estimate, where bias and variance are estimated from the posterior predictive distribution. See Keller and Rice (2017) <doi:10.1093/aje/kwx225> for more details.
README.md
eshrink
Shrinkage estimators for estimating regression parameters
This R package provides functions for estimating the penalization parameter for shrinkage estimators using the approach of Keller and Rice (2017). The package currently contains functionality for ridge regressiona and the LASSO. The penalty parameter is selected by minimizing bias and/or variance in data generated from the posterior predictive distribution.
References
Keller JP and Rice KM. (2017) Selecting Shrinkage Parameters for Effect Estimation: the Multi-Ethnic Study of Atherosclerosis. American Journal of Epidemiology. https://doi.org/10.1093/aje/kwx225