ABSTRACT
In the previous chapter, we introduced the idea of learning the parameters of a model by defining and minimising a loss function. By the end of this chapter, we will have derived exactly the same equation for the optimal parameter values from a different starting point. In particular, we will explicitly model the noise (the errors between the model and the observations) in the data by incorporating a random variable. We will demonstrate the considerable advantages of incorporating a noise term into our model. A large section of this chapter (Sections 2.2 to 2.5) is an introduction to random variables and probability which can be skipped by readers already familiar with these concepts.
