Bayesian Statistics Specialization in Python

This repository is a Python implementation of the different algorithms and problems proposed in the the courses Bayesian Statistics: From Concept to Data Analysis, Bayesian Statistics: Techniques and Models and Bayesian Statistics: Mixture Models offered by the University of California Santa Cruz in Coursera.

The main idea is to provide a Python-based implementation in order to enable people who is not familiarized with R to play and learn Bayesian Statistics.

Note: Quizes and projects are not shared in this repository, only examples given within the lectures. All the text, equations and explanations were directly taken from the course.

Table of content

The following table of content shows the different algorithms implemented in this repository.

• Course 1, Bayesian Statistics: From Concept to Data Analysis
  • L4: Plotting the Bernoully likelihood.
  • L5: Probability distributions in Python.
  • L7: Students problem with Bernoully and Binomial PMF.
  • L12: Frequentist Linear regression.
• Course 2, Bayesian Statistics: Techniques and Models
  • L3A: Markov Chains.
  • L3B: Monte Carlo.
  • L4A: Metropolis Hastings.
  • L4B: Personnel example using PyMC3.
  • L5: Gibbs sampler.
  • L6: Methods to assess convergence.
  • L7: Bayesian Linear Regression.
  • L9: Bayesian Logistic Regression.
  • L10: Poisson Regression.
  • L11A: Hierarchical modelling.
  • L11B: Hierarchical modelling - Linear Regression.
  • L11H: Mixture models.
  • LA: Custom Linear regression.
• Course 3, Bayesian Statistics: Mixture Models
  • L2: Mixture models.
  • L3: Simulating from a Mixture model.
  • L4A: Expectation-Maximization - Mixture of two Gaussians.
  • L4B: Expectation-Maximization - Mixture of three Multivariate-Gaussian.
  • L5A: Bayesian Expectation-Maximization - Mixture of two Gaussians.
  • L5B: Bayesian Expectation-Maximization - Mixture of three Multivariate-Gaussian.
  • L6: Frequentist and Bayesian density estimation with Mixture of Gaussians.
  • L7: Clustering the Iris dataset with EM and Mixture models.
  • L8: Classifying the Wine dataset with EM and Mixture models.
  • L9: Numerical stability analysis.
  • L10: Bayesian Information Criteria (BIC) for EM algorith.

Installation

To start using this repository, run the following command and install the needed dependencies.

pip install -r requirements.txt

Open jupyter notebook in your favorite IDE and enjoy!