NickGrattan (2) [Avatar] Offline
#1
Being new to R I'm attempting to execute the book's code samples in R. Some feedback on this section:

1. The term "data documentation" should state more clearly that this refers to "german.doc" on the sample data's web site.
2. Can you expand out the code section "mapping = list(" to contain all the mappings? I know that space is at a premium, but there are only another six lines to add. It makes following the code in R much easier.

I've finding the book most useful, having done quite a log of ML theory and now looking to get some practical experience. Thanks.
john.mount (79) [Avatar] Offline
#2
Re: Section 2.1.2
Thanks for your comments. Given your ideas we have decided the following.

We will try to call out the specific files by name a bit more (sounds like a good idea). I need to check later, but I think we do transform all of the columns- we just can't put that all in the book due to space. So I think we have what you want here: https://github.com/WinVector/zmPDSwR/blob/master/Statlog/GCDSteps.R .

One lesson is we need to document the structure of the zmPDSwR git store.
NickGrattan (2) [Avatar] Offline
#3
Re: Section 2.1.2
I can see from the code in GitHub there there are more than 6 extra lines I identified, so adding all to the book is not practical.

The link you provide gives all the code - the challenge then is to reference the github URLs in the book.

A single HTML page that provides a lookup from book sections to github URLs might suffice.

Thanks and regards, Nick
nina.zumel (16) [Avatar] Offline
#4
Re: Section 2.1.2
We had been planning to add a table/index to cross-reference book sections to github URLs into the book; but in addition to that a reference in the book to a single HTML page as you describe is a great idea, as the repository might change and render the (printed) book pointers obsolete.

Thanks for the suggestion!

Cheers,
Nina