Rafael-Zimmermann-s-AI-ML-Solution-for-5G-Energy-Consumption-Modelling-Challenge

Managed by Rafael Zimmermann, this repo contains the solution for the ITU AI/ML in 5G Energy Consumption Modelling Challenge.

Requirements

Python Version

Python 3.10.12 is necessary for this project.

Libraries

Install the libraries from requirements.txt: pip install -r requirements.txt

Introduction

This repo presents a comprehensive solution that takes into account three key objectives, each affecting the design and methodology of our modeling approach.

1. Objective A: Time-series forecasting methods were most effective for estimating energy consumption in specific base station products.
2. Objective B: For generalized forecasting across different but similar base stations, a hybrid model combining elements of time-series analysis and complex methods yielded the best results.
3. Objective C: Simplicity reigns supreme when generalizing across significantly different base station configurations. A simpler model ensured better performance and avoided overfitting.

Data Segmentation

Data was segmented specifically for each objective, based on features like BS_cat and RUType_cat. Masks were used to filter the test data accordingly.

Subsampling

Adversarial Validation was used for subsampling, notably for Objectives B and C, to align the training data distribution more closely with the test data.

Modeling Workflow

Common Steps for all models:

1. Data Cleaning
2. Feature Engineering
3. Ensemble Modeling: Ridge Regression + XGBoost
4. Training and Validation: MultiLabelStratifiedKFold with 10 folds

Modeling Strategy

The ensemble model merges Ridge Regression for handling linear trends with XGBoost to address non-linear patterns.

Prediction Phase

Ridge Regression provides the initial predictions, which are adjusted using XGBoost on the residuals, summing these up for the final estimates.

Best, Rafael Zimmermann