ErrorTracer
Provides a full pipeline from regularized or standard regression models (elastic net, linear models, generalized linear models, random forests) to informed Bayesian priors, structured forecast uncertainty decomposition (parameter / environmental / residual, plus a temporal component when the model carries an autocorrelation term), and forecast shelf life analysis (the quantification of when a forecast becomes uninformative). Designed for ecological and genomic forecasting with climate or environmental covariates. Methods build on Bürkner (2017) <doi:10.18637/jss.v080.i01> for Bayesian regression via 'Stan', Friedman, Hastie, and Tibshirani (2010) <doi:10.18637/jss.v033.i01> for elastic net regularization, Wright and Ziegler (2017) <doi:10.18637/jss.v077.i01> for random forests, and Vehtari, Gelman, and Gabry (2017) <doi:10.1007/s11222-016-9696-4> for leave-one-out cross-validation.
- brms: An R Package for Bayesian Multilevel Models Using Stan
- Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC
- ranger: A Fast Implementation of Random Forests for High Dimensional Data in C++ and R
- Regularization Paths for Generalized Linear Models via Coordinate Descent
This page was built for software: ErrorTracer
