Introducing Bayesian statistics

The Bayesian approach to statistics and machine learning (ML) provides a logical, transparent, and interpretable framework. This is a uniform framework that can build problem-specific models for both statistical inference and prediction. In particular, Bayesian inference offers a method to figure out unknown or unobservable quantities given known facts (observed data), employing probability to describe the uncertainty over the possible values of unknown quantities—namely, random variables of interest.

Using Bayesian statistics, we are able to express our prior assumption about unknown quantities and adjust this based on the observed data. It provides the Bayesian versions of common statistical procedures such as hypothesis testing and linear regression, covered in Chapters 11, Statistics estimation, and 12, Linear Regression in R. Compared to the frequentist approach, which we have adopted in all the models covered so far, the Bayesian approach...