Marginal likelihood of a straight-line model (eq. 28.22, Ex. 28.2, p. 352)

slope = FALSE is the horizontal-line model H1 (w1 = 0); slope = TRUE is H2 with w1 free. Closed form: t is Normal(0, priorsd^2 X X' + sigma^2 I).

Usage

morie_linevid(x, t, sigma = 1, slope = TRUE, priorsd = 1)

Arguments

x

predictor values chosen by the experimenter

t

observed responses

sigma

known noise level

slope

whether the slope is a free parameter

priorsd

prior standard deviation on each weight

Value

list(logevidence, evidence, quadform, logdet, k)

Examples

V <- c(1, 2, 3, 4, 5, 6, 7, 8)
morie_linevid(V, V)
#> $logevidence
#> [1] -10.99736
#> 
#> $evidence
#> [1] 1.674584e-05
#> 
#> $quadform
#> [1] 0.9836066
#> 
#> $logdet
#> [1] 6.308098
#> 
#> $k
#> [1] 2
#>