Metallicity RRLs

Machine learning and deep learning models to estimate metallicity from light curves.

Data

...

Dependencies

All Python packages needed are listed in requirements.txt file and can be installed simply using the pip command.

Models

The following models are implemented in this repository:

Classical ML models

1. Support Vector Regression (SVR) - A wrapper function for sklearn SVR
2. Random Forest Regressor (RF) - A wrapper function for sklearn RandomForestRegressor
3. XGBoost (XGB) - A wrapper function for XGBoost package

Deep Learning for TSER

1. Fully Convolutional Network (FCN)
2. Residual Network (ResNet)
3. Inception Time (InceptionTime)
4. LSTM
5. BiLSTM
6. ConvLSTM
7. GRU
8. BiGRU
9. ConvGRU

Output

The output folder contains all the training logs, plots, and weights, divided by model.

Utils

the folder contains all the utility files for creating models, training them, plotting the results, and saving weights and logs.

Getting started

In order to run all the experiments, the entry point file is run_experiments.py, and you can find it in the root folder.

remember to add in the file constants.py:

• The list of regressors (variable regressors).
• The dl list that identifies (Boolean) whether the regressor is of the deep learning type or not.
• the grid of the param_grid hyperparameters (model-dependent). Machine learning models have no hyperparameters here (None). Example. regressors = ["random_forest", "convgru"] dl = [False, True] e hparam_grid = [None, np.array([1, 1, 'l1', 5e-1, 5e-1, 0, 0, 0.1, 0.1])]

Support scripts

In the root folder, you can find some support scripts such as:

• pre-processing.py is necessary for the pre-processing of the photonometric light curves.
• plot_all_lightcurves.py plot of the photonometric dataset of light curves (phase/magnitude).
• evaluation_models.py uses weights from the towed model to evaluate light curves on an unknown test dataset.
• draw_models.py draws all the implemented models (found in models/ folder).

Acknowledgement

Some of the models used are based on TSER project

License

This project has an MIT-style license, as found in the LICENSE file.