Henry Garner edited this page
Mar 6, 2017
Try as we might to avoid errors creeping in, we're only human. Below are the mistakes that we've noticed, or been notified of, since going to print. Please drop us an email or tweet us if you notice any others that belong on this list.
- Page 13: It should be explicitly noted in the text that the variance and standard deviation equations presented are the population versions. Chapter 2 discusses the difference between populations and samples. (Thanks to Alex Goico).
- Page 14: xi should read xi. (Thanks to @edw.)
- Page 15: the five-number summary should be the 0, 0.25, 0.5, 0.75, and 1.0 quantiles. (Thanks to @edw.)
- Page 17:
(bin 5)instead of
- Page 39: In the section Adding columns, the phrase "Often we'll want to add rows to a dataset instead" should read "Often we'll want to add columns to a dataset instead". (Thanks to Juraj Martinka).
- Page 107:
ex-2-27returns a p-value of 0.0055, not 0.007 as printed in the book.
- Page 148: The equation for R^2 contains
(y - y-hat)in both the numerator and denominator. The denominator should be
(y - y-bar)as demonstrated the subsequent code.
- Page 162:
prediction-intervalfunction - let binding for
(/ rss dfe), not
(/ ssr dfe)as printed in the book.
- Page 201: "If you test positive, the chances of actually having the disease are 99 / 99, or 50 percent". 99 / 99 should read 99 / 198.
- Page 207:
- Page 207:
bayes-classifierfunction should take
dataas final argument. Corrected and updated code here.
- Page 214: Although the picture/not picture grouping does indeed lower the entropy of the deck, it doesn't necessarily mean it's the best question to ask. See discussion of weighted entropy and information gain on the subsequent pages. (Thanks to David Iba)
- Page 462: The creator of Minecraft is called Markus Persson, not Marcus as we incorrectly stated. We should have called him Notch like everyone else.
- Page 477: Data ordered "Jan", "Feb", "Mar" etc. is not ordered alphabetically. We know this really.