search

tapios / risk-networks

Code for risk networks: a blend of compartmental models, graphs, data assimilation and semi-supervised learning
Other
2 stars 2 forks source link

updated some codes of solvers for DA examples. #134

Closed jinlong83 closed 4 years ago

jinlong83 commented 4 years ago

Several files in 'epiforecast' folder have been modified for setting up DA examples:

  1. data_assimilator.py: resolved the issue of dividing zero when free_mass is zero;
  2. ensemble_adjustment_kalman_filter.py: added a clipping to the scaling factor of covariance when using logit transform;
  3. measurements.py: enabled users to define minimum observation noise;
  4. risk_simulator.py: added a function to set start time for the master equation solver.