makuche / multi-fidelity-bo

This repository contains data analysis scripts that I have used for my Master thesis research on multi-fidelity Bayesian optimization. The origin of the raw experimental data can be found in the thesis (not in this repository).

Multi-fidelity Bayesian Machine Learning for global optimization - data wrangling and postprocessing scripts

Project structure

Core structure of the folder:

data/                           #  Data gathered from the experiment
    data/raw/                   #  Raw data (BOSS output files)
    data/processed/             #  Pre-processed data (JSON files)
docs/                           #  Documentation of results
env/                            #  Virtual environment to process data & create plots
references/                     #  References related to the project
results/                        #  Results, including plots and tables
scripts/                        #  Scripts to process data and create plots
src/                            #  Functionalities used within the project folder
tests/                          #  Tests for the pre-processing

    flowchart TD;
    A[Raw Data - BOSS output files] -- scripts/preprocess/ --> B[Pre-processed Data ];
    B -- scripts/parse/ --> C[Data of Interest];
    C -- scripts/analyse/ --> D[Generate Plots];
    C -- scripts/analyse/ --> E[Calculate Statistics];

Loading