Deep Dive into Bayesian Modeling
Bayesian probability is a powerful technique that has revolutionized many industries by dealing with probability distributions in a different way. Discover techniques such as Hierarchical Bayesian Modeling and Markov chain Monte Carlo, and how to solve multi-armed bandits. This 2-day training offers a deep-dive into this and much more!
This Bayesian Modeling training is perfect for
- Data Scientists who know Machine Learning and want to learn about Bayesian statistics.
- This training is especially suited for Data Scientists who want to go beyond the standard probability theory.
- To get the most out of this training, we advise that you have at least one year of working experience with pandas and scikit-Learn.
What will you learn during the Bayesian Modeling training:
- You will understand what makes Bayesian Probability so powerful, especially compared to the traditional frequentist approach.
- You will learn how to use the PyMC for building Bayesian models.
- We will also teach you how about Markov Chain Monte Carlo, Variational inference, and Hierarchical Bayesian modeling.
Bayesian Probability – Program
The program of this two-day Bayesian Probability training is as follows:
Day 1: Theory and Hands-On Labs
- Fundamentals: Bayes’ Theory
- From Bayes’ Theorem to Bayesian Data Analysis
- The Bayesian’ Paradigm
- Markov chain Monte Carlo with PyMC3
- Variational Inference: Big Data Bayesian Data Analysis
- Hierarchical Bayesian Modeling with PyMC3
Day 2: Hackathon: Multi-Armed Bandits with Thompson Sampling
Multi-armed bandit problems (like for example A/B testing) can be solved by using Bayesian modeling. The most famous algorithm is Thompson Sampling, which is a Bayesian algorithm that automatically balances the exploration-exploitation tradeoff that is important in many (practical) problems in which you need to experiment to learn from the data.
During this training, you will go in-depth with Bayesian Probability. After this training, you are knowledgeable about the following topics.
You will learn:
- The theorem that underlies Bayesian data analysis and learn to apply it to solve probabilistic problems.
- How Bayes’ Theorem can be applied to data and make yourself comfortable with the Bayesian terminology: prior distributions, likelihoods, and posterior distributions.
- How The Bayesian paradigm is fundamentally different than the (more famous) “frequentist” paradigm.
- The (practical) pros and cons of working with either the Bayesian or the frequentist approach.
- Markov chain Monte Carlo (MCMC) methods
- How Variational inference offers an alternative to MCMC that is suitable to (very) big data.
This training is available in the following formats:
In-Company training is perfect for groups of 6 or more. The training takes place at your office or at one of our modern training facilities.
Online Virtual Classroom
Virtual Classrooms provide you with an interactive environment to effectively develop your skills, right from the comfort of your own home or office.