Package io.github.ai4ci.config.inhost

Interface InHostConfiguration

All Superinterfaces:
Serializable
All Known Subinterfaces:
MarkovStateModel, MarkovStateModel._PartialMarkovStateModel, PhenomenologicalModel, PhenomenologicalModel._PartialPhenomenologicalModel, StochasticModel, StochasticModel._PartialStochasticModel
All Known Implementing Classes:
ImmutableMarkovStateModel, ImmutablePhenomenologicalModel, ImmutableStochasticModel, PartialMarkovStateModel, PartialPhenomenologicalModel, PartialStochasticModel
public interface InHostConfiguration extends Serializable

  • Field Details

    • log

      static final org.slf4j.Logger log

    • LIMIT

      static final double LIMIT
      See Also:

  • Method Details

    • getInfectivityProfile

      static ImmutableDelayDistribution getInfectivityProfile(InHostConfiguration config, ExecutionConfiguration execConfig, int samples, int duration)
      Infectivity profile assumes a contact has occurred and it is the conditional probability of transmission on that day versus any other particular day. This is controlled in real life by things like symptoms and behaviour, but in theory that is controlled for by the condition that transmission has occurred. This is the difference between the generation time, and the effective generation time, and parallels R0 and Rt. This is effectively G0 not Gt, and is determined only by the average viral load in a naive host, following a standard exposure. There is a question as to whether the infectivity profile is solely dependent on viral load, or there is an element of contact behaviour in here as well. If the latter then repeated contacts would make viral load a haxard function of generation time, as people with multiple exposures eventually get infected. This would tend to make the GT look shorter than the infectious period.

    • getViralLoadProfile

      static double[] getViralLoadProfile(InHostConfiguration config, ExecutionConfiguration execConfig, int samples, int duration)
      For a configuration gets an average viral load profile. This is not a probability. A linear function of this defines the probability of transmission but this needs to be calibrated to get a population R0. The connection between viral load and infectivity profile is actually a hazard function.

    • getPeakSeverity

      static HistogramDistribution getPeakSeverity(InHostConfiguration config, ExecutionConfiguration execConfig, int samples, int duration)
      Determine the statistical distribution of maximum severity in a homogenous population, exposed with unit exposure.
      Returns:
      an empirical distribution

    • getSeveritySymptomsCutoff

      default double getSeveritySymptomsCutoff(Outbreak outbreak, ExecutionConfiguration configuration)

    • getSeverityHospitalisationCutoff

      default double getSeverityHospitalisationCutoff(Outbreak outbreak, ExecutionConfiguration configuration)
      lets say 40% asymptomatic and case hosp rate of 10%. The IHR overall is 10% of the 60% symptomatic, so 6% The cutoff is the people that don;t get hospitalised so 94% quantile.

    • getSeverityFatalityCutoff

      default double getSeverityFatalityCutoff(Outbreak outbreak, ExecutionConfiguration configuration)

    • getSeverityProfile

      static DelayDistribution getSeverityProfile(InHostConfiguration config, ExecutionConfiguration execConfig, int samples, int duration)