The Author Online Book Forums are Moving

The Author Online Book Forums will soon redirect to Manning's liveBook and liveVideo. All book forum content will migrate to liveBook's discussion forum and all video forum content will migrate to liveVideo. Log in to liveBook or liveVideo with your Manning credentials to join the discussion!

Thank you for your engagement in the AoF over the years! We look forward to offering you a more enhanced forum experience.

Gavin (41) [Avatar] Offline
#1
Perhaps mention that you should never, ever expose your TEST data any training exercises.


This evaluation is conducted in a formal way, splitting the data available into a 80% training set and 20% testing set. Another important objective is to determine if there are ay important business issues that have not been sufficiently considered.
Alessandro Negro (11) [Avatar] Offline
#2
I don't get what you mean. Give me a concrete example
Gavin (41) [Avatar] Offline
#3
Hi again,

For all my comments - please read them with an initial;
"I am happy to be wrong..."


In your statement here;

This evaluation is conducted in a formal way, splitting the data available into a 80% training set and 20% testing set. Another important objective is to determine if there are ay important business issues that have not been sufficiently considered.

I would add a sentence about the data you use for training and the data you use for testing - should never, ever be the same data.
Don't pollute one data-set with items from the other.
Not just for "this" iteration of your solution - but always.
Alessandro Negro (11) [Avatar] Offline
#4
Is this not clear in the word "split"?
Gavin (41) [Avatar] Offline
#5
Hi Alessandro,

It is clear. but I know the topic already, as do you.

But not every reader will have the appropriate base knowledge.
Only because this is the intro - "I" would be a little more specific about the requirement of them NEVER being used outside of the original tasks you assign the individual chunks of data in your 80/20 split.


Gavin.