Water level prediction using long short-term memory neural network model for a lowland river: a case study on the Tisza River, Central Europe

This repository contains code for the paper Water level prediction using long short-term memory neural network model for a lowland river: a case study on the Tisza River, Central Europe by Zsolt Vizi, Bálint Batki, Luca Rátki, Szabolcs Szalánczi, István Fehérváry, Péter Kozák and Tímea Kiss. You can find the published paper here.

Validation data

The validation data used for the analysis is located here.

Environment

The source code for this project is implemented in Python. The codes for calculations are included in the src folder and the plotting is written in Jupyter notebook.

You can download the Jupyter notebook all_analysis_for_paper.ipynb for reproducing plots in the paper from here. In this folder, you can find the data used for the visualization and the statistical analysis, but the code in the notebook downloads them automatically in the folder data.

Requirements:

• Python 3.8+
• packages listed in requirements.txt
• jupyter package to run notebooks

Structure of the project

The repository contains the following folders:

• data: the validation data and other, exported tables are downloaded to this folder via running the notebook
• notebooks: put the notebook named all_analysis_for_paper.ipynb here (all imports expect this folder as its location)
• src: Python module containing the following submodules:
  • data for data processing functionalities
  • evaluation for statistical analysis
  • model for implemented and tested models (LSTM model, Baseline, Linear and MLP)

The files containing trained weights for the models can be found in the above mentioned Google Drive folder.

Testing models

You can download the Jupyter notebook test_models.ipynb from the Google Drive folder and put into the notebooks folder to test the models implemented for this project.