Simplest example of PyMC to understand its working on a toy dataset (e.g., a coin flip model)

[muhammadalmaskhan]

Notebook proposal

Title: Simplest example of PyMC to understand its working on a toy dataset (e.g., a coin flip model)

Why should this notebook be added to pymc-examples?

This notebook aims to provide a highly accessible and transparent example for beginners to understand the fundamental structure of a PyMC model. It demonstrates a single-parameter inference (e.g., coin flip probability) using PyMC, making it an excellent starting point for newcomers.

`
import pymc as pm
import numpy as np
import matplotlib.pyplot as plt

Data: 10 coin flips with 7 heads (1 = heads, 0 = tails)

data = np.array([1, 1, 1, 1, 0, 1, 1, 0, 1, 0])

Step 1: Define a PyMC Model

with pm.Model() as model:
# Prior: Assume the probability of heads follows a Beta distribution (uniform prior, so alpha=1, beta=1)
p = pm.Beta('p', alpha=1, beta=1)

# Step 2: Define the Likelihood: This is a Bernoulli distribution based on the coin flips
obs = pm.Bernoulli('obs', p=p, observed=data)

# Step 3: Inference: Draw samples from the posterior distribution using MCMC
trace = pm.sample(2000, return_inferencedata=False)

Step 4: Visualize the posterior distribution of 'p' (the probability of heads)

pm.plot_posterior(trace, var_names=['p'])
plt.title('Posterior distribution of the probability of heads')
plt.show()
`

Suggested categories:

• Level: Beginner
• Diataxis type: Tutorial

Related notebooks

[Beginner Tutorials]: If existing beginner-level examples cover complex or multi-parameter models, this notebook will focus on a simple, single-parameter inference (e.g., Bernoulli trials for coin flips).
[Modeling Probabilistic Inference]: Can reference as a foundation to build intuition about more complex models.

References

PyMC documentation: https://www.pymc.io/projects/examples/en/latest/
Probability theory and Bayesian inference resources for beginners.