This repository contains code and data of our paper "Imprecise Bayesian Optimization". More precisely,
- bayesian-sensitivity-analysis has everything to reproduce the Sensitivity Analysis in section 4
- imprecise-bayes-opt-plug-in contains implementation of PROBO (section 6)
- imp-BO_benchmarking provides files for bechmarking experiments on graphene production data (section 7), see setup below
- univariate-benchmark-functions contains a selection of benchmark functions from R packages smoof and soobench that are used throughout the paper
- files in data allow recreating visualizations of data and functions used in the benchmark experiments, see below
- R 4.0.3
on
- Ubuntu 20.04
- Windows 10 Build 19042 (partly)
- MacOS (only visualization)
First and foremost, please clone this repo (and install required packages as indicated by your IDE)
In order to reproduce the thesis' key results for PROBO on graphene data, please
- source this file to kick off the simulation study (estimated time on 64-bit-core (linux gnu): 10h)
- results are saved automatically
- source this file to visualize the retrieved results
Or, alternatively, for PROBO on synthetic function, please
Note that it's also possible to direclty visualize the stored results by only running the visualization files. In case of own simulation, these stored results will be overwritten.
Find files to read in data and create target functions in folder data. E.g. source data/data-heartbeat-1.R to read in heartbeat time series from http://ecg.mit.edu/time-series/. Or run data/make-heartbeat-1.R to directly retrieve heartbeat target function, including a ggplot2-visualization.