Bayesian prediction
# S4 method for Bayes.fit
pred(x, invariant = FALSE, level = 0.05,
newwindow = FALSE, plot.pred = TRUE, plot.legend = TRUE, burnIn,
thinning, only.interval = TRUE, sample.length = 500,
cand.length = 100, trajectories = FALSE, ylim, xlab = "times",
ylab = "X", col = 3, lwd = 2, ...)Bayes.fit class
logical(1), if TRUE, the initial value is from the invariant distribution \(X_t~N(\alpha/\beta, \sigma^2/2\beta)\) for the OU and \(X_t~\Gamma(2\alpha/\sigma^2, \sigma^2/2\beta)\) for the CIR process, if FALSE (default) X0 is fixed from the data starting points
alpha for the predicion intervals, default 0.05
logical(1), if TRUE, a new window is opened for the plot
logical(1), if TRUE, the results are depicted grafically
logical(1), if TRUE, a legend is added to the plot
optional, if missing, the proposed value of the mixedsde.fit function is taken
optional, if missing, the proposed value of the mixedsde.fit function is taken
logical(1), if TRUE, only prediction intervals are calculated, much faster than sampling from the whole predictive distribution
number of samples to be drawn from the predictive distribution, if only.interval = FALSE
number of candidates for which the predictive density is calculated, i.e. the candidates to be drawn from
logical(1), if TRUE, only trajectories are drawn from the point estimations instead of sampling from the predictive distribution, similar to the frequentist approach
optional
optional, default 'times'
optional, default 'X'
color for the prediction intervals, default 3
linewidth for the prediction intervals, default 3
optional plot parameters
Dion, C., Hermann, S. and Samson, A. (2016). Mixedsde: a R package to fit mixed stochastic differential equations.