Enum PanjerClass 

pub enum PanjerClass {
    Binomial(Binomial),
    Poisson(Poisson),
    NegativeBinomial(NegativeBinomial),
}

A claim count in the Panjer class chosen by its dispersion Var[N] / E[N]: binomial below 1, Poisson at 1, negative binomial above 1. The usual way to state frequency uncertainty in treaty pricing (see docs/design/pareto.md).

use prospicio_prob::{Counting, PanjerClass};

let n = PanjerClass::from_mean_dispersion(4.0, 2.5).unwrap();
assert!(matches!(n, PanjerClass::NegativeBinomial(_)));
assert!((n.variance() - 10.0).abs() < 1e-12);

Variants

Binomial(Binomial)

Poisson(Poisson)

NegativeBinomial(NegativeBinomial)

Implementations

impl PanjerClass

pub fn from_mean_dispersion(mean: f64, dispersion: f64) -> Result<Self>

The claim count with this mean (finite, non-negative) and dispersion (finite, positive).

A binomial must have a whole number of trials, so for dispersion below 1 the number of trials is mean / (1 - dispersion) rounded up and p = mean / n: the mean is kept exactly and the dispersion becomes 1 - p, the nearest attainable value at or above the one asked for. PanjerClass::dispersion reports it.

pub fn dispersion(&self) -> f64

Var[N] / E[N] (1 for a zero mean).

Trait Implementations

impl Clone for PanjerClass

fn clone(&self) -> PanjerClass

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for PanjerClass

impl Counting for PanjerClass

fn pmf(&self, k: u64) -> f64

P(N = k).

fn mean(&self) -> f64

E[N].

fn variance(&self) -> f64

Var[N].

fn panjer_ab(&self) -> (f64, f64)

(a, b) with p_k = (a + b / k) p_{k-1} for k ≥ 1.

fn pgf(&self, z: f64) -> f64

Probability generating function E[z^N] for 0 ≤ z ≤ 1. Panjer’s recursion starts from P(S = 0) = pgf(f_0).

fn pgf_complex(&self, z: (f64, f64)) -> (f64, f64)

The pgf at a complex z = (re, im) with |z| ≤ 1, as (re, im). FFT aggregation applies it to the transform of the severity grid.

fn cdf(&self, k: u64) -> f64

P(N ≤ k).

fn quantile(&self, p: f64) -> Result<u64>

Smallest k with P(N ≤ k) ≥ p. p = 1 gives u64::MAX.

fn sample(&self, rng: &mut StreamRng, n: usize) -> Vec<u64>

n counts from stream rng, by inverse transform.

impl Debug for PanjerClass

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl PartialEq for PanjerClass

fn eq(&self, other: &PanjerClass) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for PanjerClass