Neural networks are used to estimate consensus hurricane track and intensity errors, as well as the associated uncertainties of the network predictions.
This code was written in python 3.9.7, tensorflow 2.7.0, tensorflow-probability 0.15.0 and numpy 1.21.4.
- Step 1:
_train_model_randomseeds.ipynb
- Step 2:
_compute_metrics.ipynb
- Step 3:
_plot_metrics.ipynb
- Step 4:
_compute_shashPredictions.ipynb
- Step 5:
_plot_singleModel.ipynb
- Step 6:
_plot_uqMetrics.ipynb
- Step 7:
_plot_uqCurves.ipynb
The following python environment was used to implement this code.
conda create --name env-hurr-tfp python=3.9
conda activate env-hurr-tfp
pip install tensorflow==2.7.0
pip install tensorflow-probability==0.15.0
pip install --upgrade numpy scipy pandas statsmodels matplotlib seaborn
pip install --upgrade palettable progressbar2 tabulate icecream flake8
pip install --upgrade keras-tuner sklearn
pip install --upgrade jupyterlab black isort jupyterlab_code_formatter
pip install silence-tensorflow
pip install tqdm
This work is a collaborative effort between Dr. Elizabeth A. Barnes and Dr. Randal J. Barnes and Dr. Mark DeMaria.
[1] Elizabeth A. Barnes and Randal J. Barnes and Nicolas Gordillo, 2021, Adding Uncertainty to Neural Network Regression Tasks in the Geosciences, arXiv 2109.07250.
This project is licensed under an MIT license.
MIT © Elizabeth A. Barnes