Machine Learning: Hands-On for Developers and Technical Professionals

Chapter 4Bayesian Networks

You might hear the Bayesian Network referred to by a few different names: probabilistic directed acyclic graphical model, Bayes Network, Belief Network, or Bayesian Model. Based on a set of variables or parameters, it's possible to predict outcomes based on probabilities. These variables are connected in such a way that the resulting value of one variable will influence the output probability of another, hence the use of networked nodes. A Bayesian Network manages to combine probability theory with graph theory and provides a very handy method for dealing with complexity and uncertainty.

This chapter covers simple Bayesian Networks and how they are used in industry. After you have mastered the simple concepts, then you can expand your study in this area.

Pilots to Paperclips

Bayesian Networks are found all over the place where uncertainty is in play, which turns out to be a lot of places. Where there is uncertainty, there is probability.

Weather forecasting and stock option predictions are examples. The financial industry uses Bayesian Networks a lot to make reasonable predictions even when the data is not complete. Bayesian Networks are the perfect tool for the likes of the insurance, banking, and investment industries. The following are a couple of specific examples of places that Bayesian Networks are being used:

• The College of Civil Aviation at Nanjing University in China has a Bayesian Network for measuring safety risk as a result of delayed ...