ABSTRACT
This chapter introduces linear regression as a Bayesian procedure. Linear regression is the geocentric model of applied statistics. Any process that adds together random values from the same distribution converges to a normal. The familiar “bell” curve of the Gaussian distribution is emerging from the randomness. Measurement errors, variations in growth, and the velocities of molecules all tend towards Gaussian distributions. The linear model strategy instructs the golem to assume that the predictor variable has a constant and additive relationship to the mean of the outcome. The golem then computes the posterior distribution of this constant relationship. Polynomial regression uses powers of a variable—squares and cubes—as extra predictors.
