lazyABC: Perform lazy ABC algorithm

In dennisprangle/lazyABCexample: Implement "Lazy ABC" paper example

Description

Perform rejection sampling lazy ABC on the SIR epidemic example or, as a special case, perform ordinary ABC

Usage

lazyABC(yobs, n.its, eps, stopstep, alpha = NULL, parallel = TRUE,
  S0 = 1e+06, I0 = 1, R0 = 0, n.subsample = 100, quiet = FALSE)

Arguments

yobs	Observed number recovered
n.its	Number of iterations to perform
eps	ABC threshold
stopstep	When to consider early stopping (set to Inf for ordinary ABC)
alpha	Continuation probability function: a function of one non-negative integer (number infectious) which returns a scalar in [0,1]. Leave NULL for ordinary ABC.
parallel	If TRUE iterations are performed in parallel
S0	Initial number susceptible
I0	Initial number infectious
R0	Initial number recovered
n.subsample	Size of subsample
quiet	Whether to suppress progress messages

Details

Prior is gamma=1 and beta~Gamma(5,1) (equivalently R0~Gamma(5,1)) Random seeds 1:n.its are used for the simulations. This is done in such a way that the ith simulation, if run to completion, will be the same regardless of alpha.

Value

A list comprising: ABCsample - dataframe of R0, weight, dist and phi; time - sum of elapsed time in each core. n.b. dist is ABC distance and phi is the decision statistic, I(stopstep).

dennisprangle/lazyABCexample documentation built on May 15, 2019, 3:25 a.m.