Multivariate double exponential (Laplace) distribution extension for SciPy

Author: David Salac https://github.com/david-salac/

Description

Generator of the random samples following the Multivariate Double Exponential (Laplace) Distribution.

Implementation is the extension of the SciPy and follows the same logic as the generator following the multivariate normal distribution.

Software User Manual (how to use script)

To generate the samples matrix, you need to have a vector of mean values and the covariance matrix. Samples are generated by the function rvs that accepts as parameters mean values vector, covariance matrix and size of generated samples.

A probability distribution can be computed using pdf function.

Example:

import numpy as np
from multivariate_laplace import multivariate_laplace

x, y = np.mgrid[-1:1:0.01, -1:1:0.01]
pos = np.empty(x.shape + (2,))
pos[:, :, 0] = x
pos[:, :, 1] = y

mean_values_vector = [0.5, -0.2]
covariance_matrix = [[2.0, 0.3], [0.3, 0.5]]

generator = multivariate_laplace(mean_values_vector, covariance_matrix)

# Get the probability distribution
probability_distribution = generator.pdf(pos)

# Get the random samples
number_of_samples = len(x)  # Redefine to what is needed
samples = multivariate_laplace.rvs(mean=mean_values_vector,
                                   cov=covariance_matrix,
                                   size=number_of_samples)

Samples are stored in the variable sample_matrix.

Run the script

If you want to run the demo script, you need to satisfy the requirements that are enumerated in the file requirements_demo.txt. You can install them directly using PIP:

pip install -r requirements_demo.txt

Default requirements

Sufficient requirements for running the multivariate Laplace distribution script is in the file requirements.txt and can be installed via PIP:

pip install -r requirements.txt