Dear Readers,
again, I tried my best to cut all the little typos, errors, and formatting bugs that slipped through the copy editing stage. Even so, while I think it is just human to have a little typo here and there, I know that this can be quite annoying as a reader!
To turn those annoyances into something positive, I will donate $5 to UNICEF USA, the US branch of the United Nations agency for raising funds to provide emergency food and healthcare for children in developing countries, for each little unreported buglet you find!
Also below, I added a small leaderboard to keep track of the errata submissions and errors you found. Please let me know if you don't want to be explicitely mentioned in that list!
- Amount for the next donation: 60$
- Amount donated: 0$
Contributor list:
- gabramson ($15)
- Gogy ($10)
- Christian Geier ($5)
- Pieter Algra / Carlos Zada ($5)
- @gabramson ($5)
- Elias Strehle ($5)
- Oliver Tomic ($5)
- Krishna Mohan ($5)
pg. 28
In the code example in the info box (sum([j * j for i, j in zip(a, b)])), it should be i * j, not ‘j * j’; otherwise, we would be calculating b^T b, not a^T b.
pg. 55
pg. 56
Wrong chapter reference in the info box:
In
Chapter 5, Compressing Data via Dimensionality Reduction{Chapter 6: Learning Best Practices for Model Evaluation and Hyperparameter Tuning}, you will learn about useful techniques, including graphical analysis such as learning curves, to detect and prevent overfitting.
pg. 91
On the top of the page, it says "Here, p (i | t ) is the proportion of the samples that belong to class c." The "c" should be changed to i.
pg. 136
The print version is incorrectly shows
>>> plt.xticks(range(X_train.shape[1]),
... feat_labels, rotation=90)instead of
>>> plt.xticks(range(X_train.shape[1]),
... feat_labels[indices], rotation=90)It seems that I did it correctly in the notebook. Also, the list of feature importances and the plot seem to be correct in the book. However, somehow the [indices] array index went missing in the print version.
pg. 155
pg. 221
There's been a mix-up, the epsilon and the 0.25 should be swapped
pg. 344
Unfortunately, there is
now{not} a universal approach for dealing with non-randomness in residual plots, and it requires experimentation
pg. 366
An error occurred so that the figure from the previous page was duplicated instead of inserting the pairwise distance matrix. Below, the correct figure is shown (note that all Jupyter Notebooks contain the correct figures):
pg. 366
The first and second columns denote the most dissimilar members in each cluster, and the third
row{column} reports the distance between those members.
pg. 371
As we can see, in this pruned clustering hierarchy, label ID_3 was
notassigned to the same cluster as ID_0 and ID_4, as expected.
pg. 506
An error during the layout ocurred so that a figure was duplicated. Below, the correct figure is shown (note that all Jupyter Notebooks contain the correct figures):
