Synthetic data set -- training and test data

This dataset contains the synthetic data generated by an agent-based disease spread model, and the results from calibration tests on that data.For full synthetic data and calibration results, download "Data.zip."For only synthetic data, download only "Training Data.zip" and "Test Data.zip." To be consistent with code, place "Training Data" and "Test Data" in a folder called "Data".<b>Contents</b>This dataset contains training and test data sets. The data is either generated in the “one-parameter case” or “two-parameter case”, where the one-parameter case is only run with one sub-population, while the two-parameter case has two sub-populations. Training data associated with calibration method 1 is generated along a discrete grid of parameter values, while training data associated with calibration method 2 is generated along continuous, randomly generated parameter values.<b>High-level organization of data</b><b>Synthetic data:</b>- Training Data - One-parameter case - Calibration method 1 - Calibration method 2 - Two-parameter case - Calibration method 1 - Calibration method 2- Test Data - One-parameter case - Two-parameter case<br><b>Calibration results:</b>- ABC_2D_Results- ABC_Results- MCMC_results - One-Pop - Two-Pop- Brute_force_posterior_estimation - One-Pop - Two-Pop<b>Data format</b>new_I_data: contains a numpy array of new infections at each time step, indexed by (simulation run, time step, sub-population). (sub-population index is dropped for datasets with only one sub-population)constant_parameters_values: csv file containing simulation information that stays constant throughout all included runs. Includes population information, including number of sub-populations, geometric centers of sub-populations, the geometric width and height (spread) of sub-populations, sub-population sizes in terms of number of agents, infection distance d_IU, fractions of sub-populations that are initially susceptible, exposed, infected, or removed/recovered, incubation and infection period information, time stepping and total time information. variable_parameter_values: csv file containing simulation information that varies between included runs. Includes mobility, jumping probability, and random seed values, indexed by run.Calibration results can be analyzed and visualized with the following codes:- ABC_2D_Results: ABC_2D_main.ipynb- ABC_Results: ABC_main.ipynb- MCMC_results/One-Pop and Brute_force_posterior_estimation/One-Pop: CalibrationMethod1_process_results.ipynb- MCMC_results/Two-Pop and Brute_force_posterior_estimation/Two-Pop: CalibrationMethod1_2D_process_results.ipynb<b>License</b>This work is licensed under CC BY 4.0