Routines to handle references constructed with the LMS method. Given a set
of LMS values, the functions convert z-scores to measurement centiles and
vice versa.
Usage
cLMS(z, L = 1, M, S)
zLMS(x, L = 1, M, S)
Arguments
z
vector of z-scores to be converted to measurements.
L
vector of Box-Cox transformation (lambda) values, L in the LMS
method.
M
vector of medians (mu), M in the LMS method.
S
vector of coefficients of variation (sigma), S in the LMS method.
x
vector of measurements to be converted to z-scores.
Value
zLMS and cLMS each return a vector or matrix,
respectively of z-scores and measurement centiles, with the number of rows
matching the length of x or z, and the number of columns
matching the length of L, M and S. If the two lengths are the same, or if
either length is 1, a vector is returned.
Details
L, M and S should all be the same length, recycled if necessary. The
formulae converting x to z and vice versa are:
$$z = \frac{(x/M)^L - 1}{L S}$$
$$x = M (1 + L S z)^{1/L})$$ where L is reset
to 10^-7 if it is zero. x and z are usually the same length as
L M and S, but can be different. The LMS method is the same as the BCCG
family in the gamlss package, except that lambda in LMS is referred
to as nu in BCCG.
# NOT RUN {cLMS(z = -2:2, L = 1:-1, M = 5:7, S = rep(0.1, 3))
cLMS(z = -2:2, L = 1:-1, M = 7, S = 0.1)
zLMS(x = 6.5, L = 1:-1, M = 5:7, S = rep(0.1, 3))
# }