Proportional data are commonly modelled using a glm
approach with logit link function. When performing the logit
transformation in advance separately, simple OLS methods can be
applied.
Usage
beta2m(b)
Arguments
b
vector or matrix holding the original data
Value
A named vector/matrix with same dimensions as b and transformed values
Details
Data are transformed according to
.
The input data are assumed to have the range 0<b<1.
Data outside this range will lead to missing values.
Corner cases (data of b=0 or b=1) can be handled by use
of fixlimits().
# NOT RUN { a <- sapply(c(0.01,0.05,0.5,0.8,0.9),function(x) rbinom(30,100,x)/100)
matplot(a,pch=20)
matplot(beta2m(a),pch=20)
matplot(a,beta2m(a),pch=20)
# }