统计学习基础(第2版)习题答案解答The Elements of Statistical Learning - solution manual.pdf

A Solution Manual and Notes for: The Elements of Statistical Learning by Jerome Friedman, Trevor Hastie, and Robert Tibshirani John L. Weatherwax? David Epstein? 16 February 2013 Introduction The Elements of Statistical Learning is an in?uential and widely studied book in the ?elds of machine learning, statistical inference, and pattern recognition. It is a standard recommended text in many graduate courses on these topics. It is also very challenging, particularly if one faces it without the support of teachers who are expert in the subject matter. For various reasons, both authors of these notes felt the need to understand the book well, and therefore to produce notes on the text when we found the text di?cult at ?rst reading, and answers to the exercises. Gaining understanding is time-consuming and intellectually demanding, so it seemed sensible to record our e?orts in LaTeX, and make it available on the web to other readers. A valuable by-product of writing up mathematical material, is that often one ?nds gaps and errors in what one has written. Now it is well-known in all branches of learning, but in particular in mathematical learning, that the way to learn is to do, rather than to read. It is all too common to read through some material, convince oneself that one understands it well, but then ?nd oneself at sea when trying to apply it in even a slightly di?erent situation. Moreover, material understood only at a shallow level is easily forgotten. It is therefore our strong recommendation that readers of the book should not look at our responses to any of the exercises before making a substantial e?ort to understand it without this aid. Any time spent in this way, even if it ends without success, will make our solutions ?wax@alum.mit.edu ?David.Epstein@warwick.ac.uk 1 easier to understand and more memorable. Quite likely you will ?nd better solutions than those provided here—let us know when you do! As far as teachers of courses based on the text are concerned