user_constraint function

Formalize User-Specified Constraint of fmx Object

Formalize user-specified constraint of fmx object

user_constraint(x, distname, K)

Arguments

• x: character vector , constraint(s) to be imposed. For example, for a two-component Tukey $g$-&-$h$

  mixture, c('g1', 'h2') indicates $g_1=h_2=0$ given $A_1 < A_2$, i.e., the $g$-parameter for the first component (with smaller location value) and the $h$-parameter for the second component (with larger mean value) are to be constrained as 0.

• distname: character scalar, name of distribution

• K: integer scalar, number of components

Returns

Function user_constraint() returns the indices of internal parameters (only applicable to Tukey's $g$-&-$h$ mixture distribution, yet) to be constrained, based on the type of distribution distname, number of components K

and a user-specified string (e.g., c('g2', 'h1')).

Examples

(d0 = fmx('GH', A = c(1,4), g = c(.2,.1), h = c(.05,.1), w = c(1,1)))
(c0 = fmx_constraint(d0))
user_constraint(distname = 'GH', K = 2L, x = character()) # equivalent

(d1 = fmx('GH', A = c(1,4), g = c(.2,0), h = c(0,.1), w = c(1,1)))
(c1 = fmx_constraint(d1))
user_constraint(distname = 'GH', K = 2L, x = c('g2', 'h1')) # equivalent

(d2 = fmx('GH', A = c(1,4), g = c(.2,0), h = c(.15,.1), w = c(1,1)))
(c2 = fmx_constraint(d2))
user_constraint(distname = 'GH', K = 2L, x = 'g2') # equivalent