Explanation of repository:

cluster/

• scripts to run notebooks
• documentation for cluster usage
• configuration for custom version of cmake
• script for downloading Kaggle data in bulk
• scripts for installing boost
• lightgbm analysis logs
• documentation for cmake
• Python library requirements specification document
• code to start a persistent jupyter lab server which is accessible from the public internet

diagrams/

draft visualisation components for the Dissertation Paper; they were transferred into vectorised graphics (see tikz/)

ensembles/

spreadsheets describing the constitution of the different multilabel classifier ensembles

lib/

code which was publicly shared by the competitors and reused in our solutions

notebooks/

the code used to define and train the various solution components; to run them, data from the Kaggle website is required

results/

spreadsheets containing summarised results from our experiments

tikz/

python code to generate vectorised graphics for LATEX