summary.abc: Summaries of posterior samples generated by ABC algortithms

Description

Calculates simple summaries of posterior samples: the minimum and maximum, the weighted mean, median, mode, and credible intervals.

Usage

# S3 method for abc
summary(object, unadj = FALSE, intvl = .95, print = TRUE,
digits = max(3, getOption("digits")-3), ...)

Value

The returned value is an object of class "table". The rows are,

Min.: minimun
Lower perc.: lower percentile
Median: or weighted median
Mean: or weighted mean
Mode: or weighted mode
Upper perc.: upper percentile
Max.: maximum

Arguments

object: an object of class "abc".
unadj: logical, if TRUE it forces to plot the unadjusted values when method is "loclinear" or "neuralnet".
intvl: size of the symmetric credible interval.
print: logical, if TRUE prints messages. Mainly for internal use.
digits: the digits to be rounded to. Can be a vector of the same length as the number of parameters, when each parameter is rounded to its corresponding digits.
...: other arguments passed to density.

Details

If method is "rejection" in the original call to abc, posterior means, medians, modes and percentiles defined by intvl, 95 by default (credible intervals) are calculated. If a regression correction was used (i.e. method is "loclinear" or "neuralnet" in the original call to abc) the weighted posterior means, medians, modes and percentiles are calculated.

To calculate the mode, parameters are passed on from density.default. Note that the posterior mode can be rather different depending on the parameters to estimate the density.

Examples

Run this code

## see ?abc for examples