Bayesian Test for Anteriority/Posteriority

A Bayesian test for checking the following assumption: "event x is older than event y".

Usage

older(x, y, ...)

# S4 method for numeric,numeric
older(x, y)

# S4 method for EventsMCMC,missing
older(x, y)

Arguments

x

A numeric vector giving the output of the MCMC algorithm for the first parameter, or an EventsMCMC object.

y

A numeric vector giving the output of the MCMC algorithm for the second parameter.

...

Currently not used.

Details

For a given output of MCMC algorithm, this function estimates the posterior probability of the event \(x < y\) by the relative frequency of the event "the value of event x is less than the value of event y" in the simulated Markov chain.

Methods (by class)

• older(x = numeric, y = numeric): Returns a length-one numeric vector (the posterior probability of the assumption: "event x is older than event y").

• older(x = EventsMCMC, y = missing): Returns a numeric matrix of posterior probabilities.

See also

Other tests: apportion(), hiatus()

Author

A. Philippe, M.-A. Vibet, N. Frerebeau

Examples

## Coerce to MCMC
eve <- as_events(events, calendar = "CE", iteration = 1)
eve <- eve[1:10000, ]

## Test for anteriority
older(eve)
#>        E1    E2     E3     E4
#> E1 0.0000 0e+00 0.4925 0.0000
#> E2 1.0000 0e+00 1.0000 0.9996
#> E3 0.5075 0e+00 0.0000 0.0000
#> E4 1.0000 4e-04 1.0000 0.0000

## Test for hiatus
hia <- hiatus(eve)
as.data.frame(hia)
#>           lower      upper  duration
#> E2-E1 -1601.050 -1059.8006 541.24908
#> E4-E1 -1062.599 -1045.3355  17.26366
#> E2-E3 -1603.779  -803.1067 800.67205
#> E4-E3 -1067.123  -793.8029 273.32032
#> E2-E4 -1599.473 -1408.9400 190.53260