dsig: Density for the Shifted Inverse Gaussian
In rettopnivek/rtclean: Outlier Removal for Response Times

Description Usage Arguments Details Value

The density function for a shifted inverse gaussian distribution, parameterized in terms of a one-boundary wiener process.

1	dsig(t, kappa, xi, tau, ln = F, summation = F)

`t`	vector of response times (t > 0).
`kappa`	a scalar threshold value (kappa > 0).
`xi`	a scalar rate of evidence accumulation (xi > 0).
`tau`	a scalar residual latency by which to shift the response times. ( 0 <= tau < min(t) ).
`ln`	logical; If TRUE, returns the log of the density.
`summation`	if TRUE, returns the sum of the logs of the densities.

The shifted inverse gaussian, when parameterized in terms of brownian motion, has density

f(t - τ) = κ/√( 2*π*(t-τ)^3 ) exp( (-.5/(t-τ))*(κ - ξ*(t-τ))^2 )

where t is a response time, κ is a threshold toward which evidence accumulates with average rate ξ and a fixed variance of 1, and τ is a residual latency subtracted from the response time.

If ln is FALSE, gives the likelihood, else gives the log-likelihood. If both ln and summation are TRUE, gives the sum of the log-likelihoods.

rettopnivek/rtclean documentation built on May 27, 2019, 5:55 a.m.