Individualized post-crisis monitoring of psychiatric patients via Hidden Markov models

This library contains the code used to generate simulation data, build the model and generate the results for the study presented in the manuscript: DOI link. There are four main parts in this library:

• Data generation: The classes in the post_crisis_monitoring/data_generation folder are used to generate the patients and simulate a hospital.
• Parameter estimation: The functions in the post_crisis_monitoring/parameter_estimation folder are used to estimate the parameters of the model presented in the manuscript.
• Model evaluation: The functions in the post_crisis_monitoring/model_evaluation folder are used to compute the expected value of the crisis.
• Plot generation: The functions in the post_crisis_monitoring/visualization folder are used to create the plots of the manuscript.

The notebook located in notebooks/generate_simulation_results.ipynb can be used to replicate the results. Additional details can be found in the manuscript: DOI link.