This class is a generic container for the model used for outbreak detection

Slots

nStates: Number of states in the hidden Markov model.
emission: Emission of hidden Markov model. Includes a excodeFamily and a excodeFormula object.
transitions: Transition proabilities of hidden Markov model.
transitions_prior: Indicates whether a prior for transition probabilities should be used.
prior_weights: Dirichlet prior weights for transition probabilities. This cannot be changed by the user.
loglik_transitions: Prior log likelihood of transition probabilities. Only relevant for model fitting.
initial_prob: Initial state probabilities of the hidden Markov model.
setBckgState: Indicates whether a background state should be computed for model fitting. Background states are set to 0s and time points where the Anscombe residuals of the initial model are < 1.
id: Name of the time series
method: Method for parameter estimation, one of c("supervised", "unsupervised"). Default: "unsupervised".
converged: Logical, indicating whether the EM-algorithm converged.
niter: Maximum number of iterations for model fitting.
LogLik: Log lokelihood of the fitted model.
AIC: Akaike information criterion of the fitted model.
BIC: Bayesian information criterion of the fitted model.
posterior: Matrix of posterior probabilities for each time point (rows) and state (columns).
alpha: Alpha (Forward) probabilities of the HMM.
pval: P-value for testing wheter timepoint is in normal or excess state.
timepoint_fit: Numeric time point in timeseries used for fitting.
timepoint: Numeric time point in timeseries used for fitting.
date: Date for each time point.
observed: Vector of observed number of cases.
population: Population size, default: 1.
error: Character string of error message (if an error occured).