In the previous chapters, we looked at the different types of models and how to create models for our problems. We also saw how the probabilities of variables change when we change the probabilities of some other variables. In this chapter, we will be discussing the various algorithms that can be used to compute these changes in the probabilities. We will also see how to use these inference algorithms to predict the values of variables of new data points based on our model, which was trained using our previous data.
In this chapter, we will cover:
Using inference to answer queries about the model
Variable elimination
Understanding the belief propagation algorithm using a clique tree
MAP inference using variable elimination
MAP inference using belief propagation
Comparison between variable elimination and belief propagation