Software: Adaptive Approximate Bayesian Computation
A Julia package stored in a GitHub repository that contains an Approximate Bayesian Computation (ABC) model with distance functions. This is used as a likelihood-free approach to calibrating a given simulation function to some observed data.
There are 2 ABC methods available, rejection-ABC and Population Monte Carlo ABC (PMC-ABC), each with 2 different distance functions, the block bootstrap distance function and the weighted euclidean distance function.
This package is used in the dissertation, Calibrating a Latent Order Book Model to Market Data, to calibrate various models (ARMA, LOB, SLOB) to synthetic and real market data.