AdaptiveABC

Installation

To install the package under Julia 0.7+ using Pkg, simply enter its interactive mode by pressing ] in the Julia REPL and run

add https://github.com/istvankleijn/AdaptiveABC.jl

After this, your Julia install will track AdaptiveABC's master branch. Installing updates works the same way as for registered packages, i.e. update in the Pkg interactive mode. See also https://docs.julialang.org/en/v1.0/stdlib/Pkg/ for more details.

Most important changes in this fork

Although the file structure has changed from https://github.com/aifbowman/ABC-SMC, the original APMC() function still houses the ABC-APMC algorithm and is called using the same arguments. However, it now returns a APMCResult type, with rather different field names from before. See https://github.com/istvankleijn/AdaptiveABC.jl/blob/master/src/types.jl for the type definitions.

Furthermore, the package includes a wrapper modelselection() function. Currently, this takes an APMCInput type that differs primarily from the input arguments to APMC() in requiring separate data-generating functions (simulators) and distance functions (metrics). In future updates, I aim to make the modelselection() interface more general (e.g. allowing for rejection-ABC only, or continuing a previous result) and add a parameterinference() wrapper.