FIDE

Data

All data files can be found in the Data folder.

Code

All code files are available in the Code folder (todo: Example notebook will be uploaded).

Code Files

• data_process.py: Contains data processing steps.
• distribution_evaluation.py: Runs all distribution evaluation functions (defined in general_utilities).
• general_utilities.py: Contains general functions and distribution evaluation functions.
• evaluation_predictive_discriminative_score_all_values.py: Contains evaluations of predictive and discriminative scores (according to TimeGAN paper) for all time step values.
• evaluation_predicitive_score_block_maxima.py: Contains evaluations of predictive scores (according to TimeGAN paper) for only block maxima.
• main_temperature_example.py: Contains the main program with example temperature data.
• model.py: Contains the conditional diffusion-based model.
• sampling.py: Contains the sampling procedure for generating samples.
• train_utilities.py: Contains functions for training, including plotting and metrics calculations.
• TSNE_evaluation.py: Measures and plots t-SNE of the real and generated data.