gridsearch is an R package used for adjudicating plausible data generation processes (DGP) for a given panel dataset. In its current version, gridsearch allows researchers to:
- Generate simulated datasets from a given set of DGP parameters,
- Implement a simple Approximate Bayesian Computation (ABC) algorithm to see what plausible combinations of parameter values may approximate to the observed data,
- Plot posterior and conditional distributions generated from the grid-search algorithm.
This package is a companion to the article, “The Promises and Pitfalls of Using Panel Data to Understand Individual Belief Change,” stored in SocArXiv.
Installation
You can install the development version of gridsearch from Github with:
# install.packages("devtools")
devtools::install_github("tkeskinturk/gridsearch")