Skip to contents

Three-parameter Burr Type XII distribution with scale \(s\), first shape \(\zeta\) (shape1), and second shape \(\theta\) (shape2, the tail index). The distribution is supported on \(x > 0\) and is widely used for modelling heavy-tailed positive variables. L-moments are computed via closed-form expressions involving Beta-function PWMs; a two-step seeding procedure initialises the shapes from a pre-computed lookup table (Burr_InitValues) and derives the scale analytically. Fitting uses L-BFGS-B minimisation of the normalised L-moment error.

The probability density function is: $$f(x) = \zeta s^{-\zeta} x^{\zeta-1} \left(\zeta\theta\left(\frac{x}{s}\right)^\zeta + 1\right)^{-1/(\zeta\theta) - 1}, \quad x > 0$$ where:

  • \(s\) — scale parameter (scale)

  • \(\zeta\) — first shape parameter (shape1), controls lower-tail behaviour

  • \(\theta\) — second shape parameter (shape2), tail index; the mean exists only when \(\theta < 1\)

Usage

dburr(x, scale, shape1, shape2, PW = 1)

pburr(q, scale, shape1, shape2, PW = 1)

qburr(p, scale, shape1, shape2, PW = 1)

rburr(n, scale, shape1, shape2, PW = 1)

Arguments

x, q

vector of quantiles.

scale

scale parameter.

shape1

first shape parameter.

shape2

second shape parameter.

PW

probability weight (point mass at zero for zero-inflated variants).

p

vector of probabilities.

n

number of observations. If length(n) > 1, the length is taken as the number required.