The Elements of Statistical Learning

Trevor Hastie, Robert Tibshirani and Jerome Friedman

The learning problems that we consider can be roughly categorized as either supervised or unsupervised. In supervised learning, the goal is to predict the value of an outcome measure based on a number of input measures; in unsupervised learning, there is no outcome measure, and the goal is to describe the associations and patterns among a set of input measures.


There is no true interpretation of anything; interpretation is a vehicle in the service of human comprehension. The value of interpretation is in enabling others to fruitfully think about an idea.   —Andreas Buja  


We have tried to write this book in an intuitive fashion, emphasizing concepts rather than mathematical details.


Our view is that one must understand simple methods before trying to grasp more complex ones.


We recommend that Chapters 1—4 be first read in sequence. Chapter 7 should also be considered mandatory, as it covers central concepts that pertain to all learning methods. With this in mind, the rest of the book can be read sequentially, or sampled, depending on the reader’s interest.


The symbol indicates a technically difficult section, one that can be skipped without interrupting the flow of the discussion.


This distinction in output type has led to a naming convention for the prediction tasks: regression when we predict quantitative outputs, and classification when we predict qualitative outputs. We will see that these two tasks have a lot in common, and in particular both can be viewed as a task in function approximation.


The linear decision boundary from least squares is very smooth, and apparently stable to fit. It does appear to rely heavily on the assumption that a linear decision boundary is appropriate.