Machine learning: Bayesian Networks

One 、 What is Bayesian network

Bayesian network is a model for probabilistic reasoning .（ For example, the following figure , Arrows indicate causality , That is, robbers and earthquakes will cause the house bell to ring , If the house bell rings , Then this man's two neighbors John and mary Will call him ）. The network established through causality here is called Bayesian Network , So what reasoning does it support ？

If we build this Bayesian network based on prior knowledge , Then we can carry out probabilistic reasoning on such a query —— If John Call me up , What is the probability of robbery ？

This process of probabilistic reasoning applies Bayesian formula , So we call it Bayesian Network .

Two 、 How to build Bayesian Networks ？

In fact, this is very trivial Of , Bayesian networks are constructed by domain experts , All prior probabilities are based on empirical statistics . If the bell rings, Bayesian networks can be built by people with common sense . Prior knowledge ,e.g P(E) It is the probability of statistics based on the number of earthquakes in the past （ Others also get ）.

---

---

---

Reference material ：
machine learning ： Bayesian network