README.md

Experimental study of time series forecasting methods for groundwater level prediction

This repository contains the source code and the results of our work on comparing global vs local forecasting method for groundwater level prediction.

Folders description

• dataset_collection: describes how the data is collected and provides helping scripts for downloading the raw data.
• local_global_experiments: Contains scripts for running the models on the datasets.
• portfolio_experiments: Contains a Jupyter notebook for visualising the datasets.
• results: Contains the result files and a Jupyter notebook for visualising the results in a human-friendly way.

Download the the formatted dataset

Dependencies

There are conda and pip dependencies respectively listed in the files environment.yml and requirements.txt. Both can be installed simultaneously using the following command:
conda env create -f=environment.yml

Some results

• Some forecasts with NeuralProphet
• RMSSE
• Detailed results in CSV

Citing this work

Bibtex is coming...