ParaMonte Fortran 2.0.0
Parallel Monte Carlo and Machine Learning Library
See the latest version documentation.
pm_distPiwiPoweto Module Reference

This module contains classes and procedures for computing various statistical quantities related to the (Truncated) PieceWise Power/Pareto distribution (hence the name PiwiPoweto). More...

Data Types

interface  getPiwiPowetoCDF
 Generate and return the Cumulative Distribution Function (CDF) of the (Truncated) PiwiPoweto distribution for an input logx within the support of the distribution logLimX(1) <= logx <= logLimX(size(logLimX)).
More...
 
interface  getPiwiPowetoLogPDF
 Generate and return the natural logarithm of the Probability Density Function (PDF) of the (Truncated) PiwiPoweto distribution for an input logx within the support of the distribution logLimX(1) <= logx <= logLimX(n+1).
More...
 
interface  getPiwiPowetoLogPDFNF
 Generate and return the natural logarithm of the normalization factors of the components of the Probability Density Function (PDF) of the (Truncated) PiwiPoweto distribution for the input parameter vectors \((\alpha, x_\mathrm{lim})\). More...
 
interface  setPiwiPowetoCDF
 Return the Cumulative Distribution Function (CDF) of the (Truncated) PiwiPoweto distribution for an input logx within the support of the distribution logLimX(1) <= logx <= logLimX(size(logLimX)).
More...
 
interface  setPiwiPowetoLogPDF
 Return the natural logarithm of the Probability Density Function (PDF) of the (Truncated) PiwiPoweto distribution for an input logx within the support of the distribution logLimX(1) <= logx <= logLimX(n+1). More...
 

Variables

character(*, SK), parameter MODULE_NAME = "@pm_distPiwiPoweto"
 

Detailed Description

This module contains classes and procedures for computing various statistical quantities related to the (Truncated) PieceWise Power/Pareto distribution (hence the name PiwiPoweto).

Specifically, this module contains routines for computing the following quantities of the (Truncated) PieceWise Power/Pareto distribution:

  1. the Probability Density Function (PDF)
  2. the Cumulative Distribution Function (CDF)
  3. the Random Number Generation from the distribution (RNG)
  4. the Inverse Cumulative Distribution Function (ICDF) or the Quantile Function

The (Truncated) PiwiPoweto distribution is better known by its special case: the Broken Power Law distribution.

The PDF of an \(n\)-piece (Truncated) PiwiPoweto distribution (with \(n > 0\)) over a strictly-positive support \(x \in [0 < x_\mathrm{lim,1}, x_\mathrm{lim,n+1} \leq +\infty)\) is defined with the ascending-ordered scale vector \(x_\mathrm{lim} = \{ x_\mathrm{lim,i}, i = 1 : n + 1\}\) and shape vector \(\alpha = \{\alpha_i, i = 1 : n \}\) as,

\begin{equation} \large \pi(x | \alpha, x_\mathrm{lim}) = \begin{cases} \eta_1 x^{\alpha_1 - 1} &, ~ 0 \leq x_\mathrm{lim,1} \leq x < x_\mathrm{lim,2} ~, \\ \ldots \\ \eta_i x^{\alpha_i - 1} &, ~ x_\mathrm{lim,i} \leq x < x_\mathrm{lim,i+1} ~, \\ \ldots \\ \eta_n x^{\alpha_n - 1} &, ~ x_\mathrm{lim,n} \leq x < x_\mathrm{lim,n+1} \leq +\infty ~, \end{cases} \end{equation}

where the conditions \(\{\alpha_i \in \mathbb{R}, i = 1 : n\}\) and \(\alpha_1 > 0 \lor x_\mathrm{lim,1} > 0\) and \(\alpha_n < 0 \lor x_\mathrm{lim,n+1} < +\infty\) must hold.
These conditions must be met for the PDF to be normalizable.

The component normalization factors \(\eta_i\) are computed according to the following relationship,

\begin{equation} \large \eta_i = \eta_1 \prod_{j = 2}^i ~ x_\mathrm{lim,j}^{(\alpha_{j - 1} - \alpha_j)} ~, ~i = 2:n \end{equation}

where \(\eta_1\) is a normalization factor that properly normalizes the integral of the PDF over its support to unity.

The corresponding Cumulative Distribution Function (CDF) of the (Truncated) PiwiPoweto is,

\begin{equation} \large \mathrm{CDF}(x | \alpha, x_\mathrm{lim}) = \sum_{i = 1}^{n : x_\mathrm{min,n} < x} ~ S_i ~, \end{equation}

where \(S_i ~:~ 1 \leq i \leq n\) is expressed via the following,

\begin{equation} \large S_i = \begin{cases} \frac{\eta_i}{\alpha_i} \log\big( \frac{\min(x, x_{i+1})}{x_i} \big) &, ~\alpha_i = 0 ~, \\ \frac{\eta_i}{\alpha_i} \big( \min(x, x_{i+1})^{\alpha_i} - x_{i}^{\alpha_i} \big) &, ~\alpha_i \neq 0 ~, \end{cases} \end{equation}

where \(\eta_i\) are the normalization factors of the Power-Law components of the distribution.

See also
pm_distPower
pm_distPareto
Test:
test_pm_distPiwiPoweto
Todo:
Generic interfaces for computing the logarithm of CDF robustly (without numerical rounding) must be added in the future.


Final Remarks


If you believe this algorithm or its documentation can be improved, we appreciate your contribution and help to edit this page's documentation and source file on GitHub.
For details on the naming abbreviations, see this page.
For details on the naming conventions, see this page.
This software is distributed under the MIT license with additional terms outlined below.

  1. If you use any parts or concepts from this library to any extent, please acknowledge the usage by citing the relevant publications of the ParaMonte library.
  2. If you regenerate any parts/ideas from this library in a programming environment other than those currently supported by this ParaMonte library (i.e., other than C, C++, Fortran, MATLAB, Python, R), please also ask the end users to cite this original ParaMonte library.

This software is available to the public under a highly permissive license.
Help us justify its continued development and maintenance by acknowledging its benefit to society, distributing it, and contributing to it.

Author:
Amir Shahmoradi, Oct 16, 2009, 11:14 AM, Michigan

Variable Documentation

◆ MODULE_NAME

character(*, SK), parameter pm_distPiwiPoweto::MODULE_NAME = "@pm_distPiwiPoweto"

Definition at line 100 of file pm_distPiwiPoweto.F90.