With the goal to have an "outlier"-aware normal distribution, I build a simple univariate mixing model of normal and uniform distributions:
U1 <- distr::Unif(-100, -4) U2 <- distr::Unif(4, 100) N <- distr::Truncate(distr::Norm(), lower = -4, upper = 4) C <- distr::UnivarMixingDistribution(U1, N, U2, mixCoeff = c(1, 98, 1)/100)
However, the quantile function is not the inverse of the distribution function:
x <- distr::p(C)(0.25) x > 0.5967383 distr::q(C)(x) > 0.2500331
Is this inaccuracy caused by my inaccuracte mixing model or by a floating point issue?