Thoughtful Machine Learning

Errata for Thoughtful Machine Learning

Submit your own errata for this product.


The errata list is a list of errors and their corrections that were found after the product was released. If the error was corrected in a later version or reprint the date of the correction will be displayed in the column titled "Date Corrected".

The following errata were submitted by our customers and approved as valid errors by the author or editor.

Color Key: Serious Technical Mistake Minor Technical Mistake Language or formatting error Typo Question Note Update



Version Location Description Submitted By Date Submitted Date Corrected
PDF
Page 200
Table 10-5

The text states that the precision of Stout is 30/36 but the table suggests that it would be 30/32.

Note from the Author or Editor:
Agreed. It should be 30/32 which would suggest a precision of 93.75% This means change the sentence to "For instance, the precision of Stout is 30/32, or around 93%. The recall of stout is also 30/32"

Brian Drye  Nov 05, 2014  Jan 16, 2015
PDF
Page 170
Figure 9-1

Graphically collaborative filtering looks something like Figure 9-1 (where beer has been substituted for tea). There's no tea in the figure. Also, the figure doesn't add any information beyond what is stated in the text.

Note from the Author or Editor:
Oh boy... Yea Originally this example was tea. It should only be beer. Take out "(where beer has been substituted for tea)"

Brian Drye  Nov 04, 2014  Jan 16, 2015
PDF
Page 139
3rd paragraph

Inconsistent usage of "Neural Networks" and "Neural Nets"

Note from the Author or Editor:
Should be Neural Networks instead of Neural Nets

Brian Drye  Nov 04, 2014  Jan 16, 2015
PDF
Page 131
2nd paragraph

Confusing sentence & grammar: "The periodic functions (shown in Figures 7-7 and 7-9) are used for data that is random and is the cosine and sine." What is the subject that corresponds to "is the cosine and sine"?

Note from the Author or Editor:
Yes this is a typo after editing too rigorously. Should read "The periodic functions (shown in Figures 7-7 and 7-9) are used for modeling data with lots of noise. They generally take the form of either a sine or cosine function"

Brian Drye  Nov 04, 2014  Jan 16, 2015
PDF
Page 77
Figure 5-2

State machine numbers do not match the numbers listed in Table 5-1. Examples: User to Prospect should be 0.05 (not 0.5) User to Customer should be 0.15 (not 0.05) Customer to Customer should be 0.03 (not 0.3) Confusing when the probabilities do not add up to one for each state.

Note from the Author or Editor:
Yes thank you this is correct

Brian Drye  Nov 03, 2014  Jan 16, 2015
Printed, PDF, ePub
Page 54
4th paragraph

Confusing: If there are 1000 orders, with: 40 fraudulent without gift cards 60 fraudulent with gift cards 40 not fraudulent with gift cards If you look at all the gift card orders (100), you will find 60 fraudulent orders. The last sentence reads: "Out of the remaining 900, 90 used gift cards, which brings the total we need to look at to 150!" This isn't correct. We just looked at all the gift card orders, so out of the 900 remaining orders, none have gift cards. By looking at the gift card orders, you will find 60 fraudulent orders, but there will still be 40 fraudulent orders somewhere in the remaining 900.

Note from the Author or Editor:
This one is confusing I agree. Basically the idea is that we have 1000 orders. 100 of the orders in the past have been fraudulent. So 10% of overall orders. Out of the fraudulent orders the probability of it having a gift card is 60% therefore we have 60 orders inside of the fraudulent orders that are gift card orders. From there we have 900 orders that are not fraudulent. So that's 90% of all orders, in this case though the probability of them having a gift card is less (10%). So therefore there are 90 non-fraudulent gift card orders. You add the two together and that's 150 gift card orders. The thing I didn't clarify here though is that we're ignoring the other 40 fraudulent orders cause we don't have a viable signal to find them. Does that make sense?

bdrye  Nov 02, 2014  Jan 16, 2015
Printed, PDF, ePub
Page 51
Heading, and elsewhere

Grammar: apostrophe does not need a 's' "...Bayes's Theorem..." should be "...Bayes' Theorem..." Ex: Using Bayes’s Theorem to Find Fraudulent Orders

Note from the Author or Editor:
I agree the general way to write Bayes' theorem is with one apostrophe.

bdrye  Nov 02, 2014  Jan 16, 2015
PDF
Page 53
Figure 4-1

the first P(B|A) should be P(A|B). P(A|B) = 6%/20% = 30% P(B|A) = 6%/30% = 20%

Note from the Author or Editor:
Yup confirmed.

bdrye  Nov 02, 2014  Jan 16, 2015
Printed, PDF, ePub
Page 28
Figure 3-5

"Lawful (1)" should be "Lawful (2)"

Note from the Author or Editor:
Yes agreed. In this example Neutral (0) should be Neutral (1)

Brian Drye  Oct 28, 2014  Jan 16, 2015
Printed, PDF, ePub
Page 27
Figure 3-3

"Bad(3)" should be "Bad(2)"

Note from the Author or Editor:
Yes Bad(3) should be labeled as Bad(2)

Brian Drye  Oct 28, 2014  Jan 16, 2015
PDF
Page 19
top

Table 2-3, "Mathematical Notation used throughout the book," has some LaTeX formatting problems.

Note from the Author or Editor:
Checked with the author and he sent me edits.

Dan Bernier  Sep 12, 2014  Sep 22, 2014