How to Use the Log-Normal Distribution in Python

Introduction

The log-normal distribution is a continuous probability distribution of a random variable whose logarithm is normally distributed. In other words, if the random variable X is log-normally distributed, then Y = ln(X) has a normal distribution. This distribution is applicable to various fields, including finance, hydrology, insurance, and many other disciplines.

Python, a dynamic and versatile programming language, provides several libraries that make it easy to work with different statistical distributions, including the log-normal distribution. In this article, we’ll delve into the concept of the log-normal distribution and illustrate how to implement it using Python.

Understanding the Log-Normal Distribution

A log-normal distribution is characterized by two parameters: the mean (mu, μ) and the standard deviation (sigma, σ) of the variable’s natural logarithm. Notably, these are not the mean and standard deviation of the variable itself.

If a random variable X follows a log-normal distribution, it will take only positive real values, and its distribution will be right-skewed. The log-normal distribution is used in various fields, such as in modeling stock prices or housing prices, where values cannot go below zero and can potentially go very high.

Generating a Log-Normal Distribution with NumPy

NumPy is a fundamental library for scientific computing in Python. It offers a function numpy.random.lognormal to generate a log-normal distribution. This function requires two parameters: the mean (mu) and standard deviation (sigma) of the underlying normal distribution. It also takes an optional size parameter to specify the shape of the returned array.

Here’s how to generate a log-normal distribution with a mean of 0 and a standard deviation of 1:

import numpy as np

mu = 0  # mean
sigma = 1  # standard deviation
size = 1000  # output shape

# Generate log-normal distribution
distribution = np.random.lognormal(mu, sigma, size)

print(distribution)

The output will be an array of 1000 numbers, each drawn from a log-normal distribution.

Visualizing a Log-Normal Distribution with Matplotlib

We can use Python’s Matplotlib library to create a histogram and visualize our distribution. Here’s how:

import matplotlib.pyplot as plt

plt.hist(distribution, bins=100, density=True)
plt.xlabel('Value')
plt.ylabel('Frequency')
plt.title('Log-Normal Distribution (mu=0, sigma=1)')
plt.show()

The ‘density=True’ parameter ensures the histogram is a probability density plot, which means the area under the histogram will integrate to 1.

Analyzing a Log-Normal Distribution with SciPy

The SciPy library is another fundamental library for scientific computing in Python. It provides the scipy.stats.lognorm object, which represents a log-normal distribution and has methods for calculating properties of this distribution, such as the probability density function (pdf), cumulative distribution function (cdf), and more.

For example, we can calculate the probability density of certain values:

from scipy.stats import lognorm

# Calculate probability density
pdf_values = lognorm.pdf(distribution, sigma, scale=np.exp(mu))

print(pdf_values)

And the cumulative probability up to certain values:

# Calculate cumulative probability
cdf_values = lognorm.cdf(distribution, sigma, scale=np.exp(mu))

print(cdf_values)

Here, ‘scale’ is set to np.exp(mu) because lognorm in SciPy parameterizes the distribution by sigma and scale, where scale is equal to exp(mu).

Conclusion

Understanding the log-normal distribution is fundamental in many areas of study, especially those dealing with non-negative data and with a large range of values. Python and its libraries, such as NumPy, Matplotlib, and SciPy, make it easier to work with this distribution, providing functions to generate random variates, visualize the distribution, and calculate important properties.

As you continue your journey in data science or any field requiring statistical analysis, you’ll find Python’s extensive capabilities and robust libraries to be valuable allies. Keep practicing and exploring more, and you’ll be well-equipped to tackle any statistical challenge you encounter.