This repository contains code submitted for the United Nations Economic Commission for Europe (UNECE) HLG-MOS, WP1: Pilot Study, Classification and Coding.
The data used in this repository is open-source ECOICOP classified products provided by Statistics Poland.
Source: https://github.com/UNECE/ML_dataset.
The repository is a series of Juypter Notebook files which break down a machine-learning pipeline into four steps: preprocessing (Step1), hyperparameter tuning (Step2), training the model (Step3), and evaluating the model (Step4).