Is the keras model in adult sensus data realy is the best?

[kyxaxa]

page 141: "you will notice that this approach gives the best results".
But only in this approach you did not show the results.

I've got np.average(aucs)=0.893 and this is the worst result, not the best result.
And I don't know - that was my mistake (some error in code) or something else.

[shun74]

I think you are misunderstanding now.

I also thought the score in last script isn't the best, because post script "target_encoding.py" scored almost 0.93 .

But actually it's the best score, because datasets used in these two scripts are different.

The author compare to script in 19 or 20 pages before the last script.

I know that it's very confusing at first. I hope you can understand it.