Linear regression
There are many different linear regression models built-in in Scikit-learn, Ordinary Least Squares (OLS) and Least Absolute Shrinkage and Selection Operator (LASSO) to name two. The difference between these two can be approximated by different loss functions, which is the function that is worked on by the machine learning algorithm. In LASSO, there is an added penalty going away from the fitted function, whereas OLS is simply the least square equation. However, the routine is still different from the OLS that we covered earlier; the underlying algorithm to reach the answer is a machine learning algorithm. One such common algorithm is gradient decent. Here, we shall take the climate data from the previous chapter and fit a linear function to it with two methods, then we will compare the results from the OLS model with that of PyMC's Bayesian inference ( Chapter 6 , Bayesian Methods) and statsmodels' OLS ( Chapter 4 , Regression).
Climate data
We begin by reading in the...
