Link functions in INLA

Define link-functions and its inverse

inla.link.cauchit(x, inverse = FALSE)

inla.link.invcauchit(x, inverse = FALSE)

inla.link.log(x, inverse = FALSE)

inla.link.invlog(x, inverse = FALSE)

inla.link.neglog(x, inverse = FALSE)

inla.link.invneglog(x, inverse = FALSE)

inla.link.logit(x, inverse = FALSE)

inla.link.invlogit(x, inverse = FALSE)

inla.link.probit(x, inverse = FALSE)

inla.link.invprobit(x, inverse = FALSE)

inla.link.robit(x, df = 7, inverse = FALSE)

inla.link.invrobit(x, df = 7, inverse = FALSE)

inla.link.loglog(x, inverse = FALSE)

inla.link.invloglog(x, inverse = FALSE)

inla.link.cloglog(x, inverse = FALSE)

inla.link.invcloglog(x, inverse = FALSE)

inla.link.ccloglog(x, inverse = FALSE)

inla.link.invccloglog(x, inverse = FALSE)

inla.link.tan(x, inverse = FALSE)

inla.link.invtan(x, inverse = FALSE)

inla.link.tan.pi(x, inverse = FALSE)

inla.link.invtan.pi(x, inverse = FALSE)

inla.link.identity(x, inverse = FALSE)

inla.link.invidentity(x, inverse = FALSE)

inla.link.inverse(x, inverse = FALSE)

inla.link.invinverse(x, inverse = FALSE)

inla.link.invqpoisson(x, inverse = FALSE, quantile = 0.5)

inla.link.sn(x, intercept = 0.5, skew = 0, a = NULL, inverse = FALSE)

inla.link.invsn(x, intercept = 0.5, skew = 0, a = NULL, inverse = FALSE)

inla.link.gevit(x, tail = 0.1, inverse = FALSE)

inla.link.invgevit(x, tail = 0.1, inverse = FALSE)

inla.link.cgevit(x, tail = 0.1, inverse = FALSE)

inla.link.invcgevit(x, tail = 0.1, inverse = FALSE)

inla.link.invalid(x, inverse = FALSE)

inla.link.invinvalid(x, inverse = FALSE)

Arguments

x

The argument. A numeric vector.

inverse

Logical. Use the link (inverse=FALSE) or its inverse (inverse=TRUE)

df

The degrees of freedom for the Student-t

quantile

The quantile level for quantile links

intercept

The quantile level for the intercept in the Skew-Normal link

skew

The skewness in the Skew-Normal. Only one of skew and a can be given.

a

The a-parameter in the Skew-Normal. Only one of skew and a can be given.

tail

The tail parameter in the GEV distribution (0 < tail <= 1/2)

Value

Return the values of the link-function or its inverse.

Note

The inv-functions are redundant, as inla.link.invlog(x) = inla.link.log(x, inverse=TRUE) and so on, but they are simpler to use as arguments to other functions.

Author

Havard Rue hrue@r-inla.org